# Generate a text SOP from a recording
Canonical URL: /10x-sop-pilot
Description: Turn a project made from one of your recordings into a standardized, text-only standard operating procedure that contains only what the recording shows.
This is an early beta, on by request for your workspace. The goal of this round
is to run it on your real recordings and tell us where the output needs work.
It writes only what your recording shows or says. It does not add roles,
schedules, policies, or links that the recording did not establish. Anything the
recording does not cover is left as a visible gap for your team to fill, never
guessed.
## What you get [#what-you-get]
The command produces a standalone, text-only SOP with a fixed set of 11
headings:
1. SOP Title
2. Purpose
3. Owner / Responsible Role
4. When / Why This SOP Is Used
5. Frequency / Timing
6. Required Access or Prerequisites
7. `Step-by-Step Instructions`
8. Important Notes or Exceptions
9. Escalation / Troubleshooting
10. Expected Outcome / Completion Criteria
11. Related Resources / Links
## Before you start [#before-you-start]
You need a project that was created from a recording, such as a Loom or a screen
capture your team imported into Layerpath. Open that project in the editor.
## How to run it [#how-to-run-it]
1. Open the project, then open Agent from the toolbar at the bottom of the
editor.
2. In the Agent box, type `/10x-sop` and send it. A plain request such as
"convert this into a 10X SOP" also works.
3. Wait while it runs. A short project takes under a minute; a longer one can
take two to three minutes. Keep the tab open while it runs.
4. Review the result and use Copy SOP to paste it wherever your team keeps
documentation.
## How to read the result [#how-to-read-the-result]
* Gaps are intentional. Where the recording did not establish something, that
heading reads "Not specified in recording." It is a prompt for your team to
fill in, and it means the tool chose not to invent an answer.
* Related screen actions are grouped into logical steps that keep the detail,
rather than one line per click.
* Named tools and spreadsheets are carried through when the recording names
them, with no invented links.
## Good to know [#good-to-know]
* It does not change your project. The SOP is generated output, not an edit.
* The result is not saved. Copy it out before you close the panel; refreshing
the editor clears it.
* If you see "A 10X text SOP is already being generated for this project," a run
is still in progress. Wait a couple of minutes and try again rather than
sending it more than once.
## Rough edges we already know about [#rough-edges-we-already-know-about]
Because this is a beta, expect light cleanup on some outputs, usually under ten
minutes:
* A heading such as Purpose may read "Not specified in recording." when the
wording could not be tied to the recording. That is the no-invention rule
being strict.
* Purpose and When / Why This SOP Is Used can read similarly on some projects.
* A tool or spreadsheet name may appear slightly differently in one place than
another.
None of these add invented content. If anything reads as made up, that is what
we most want to hear about.
## What would help us most [#what-would-help-us-most]
* Run it on a handful of your existing recordings.
* Tell us, per SOP, whether you could use it without watching the recording, and
roughly how much cleanup it needed.
* Flag anything that reads as invented, or any heading that came out wrong.
---
# Layerpath documentation
Canonical URL: /
Description: Learn how to create a project, edit its tour and steps, and present it as an interactive tour, guide, or video export.
# Create a project. Share it as an interactive tour, guide, or video export. [#create-a-project-share-it-as-an-interactive-tour-guide-or-video-export]
Start with existing material or record a browser workflow. Then improve it and share it in the format your audience needs.
Create your first project
## How should I start? [#how-should-i-start]
Choose the path that matches the material you have now.
Add new images or video, reuse library media, or use a Loom or YouTube link, then describe what viewers should learn.
Capture a browser task in the order viewers should follow it.
## What happens next? [#what-happens-next]
Review the initial tour, make focused changes, then preview the complete experience.
Confirm that the project is ready before judging its final step order and content.
Review the ordered steps and make changes in the editor.
Decide whether viewers should use an interactive tour, read a guide, or watch a video export.
## How Layerpath fits together [#how-layerpath-fits-together]
A **workspace** contains projects. Each project has one editable tour made of ordered steps. The tour is the source for the interactive experience, guide, and video export.
[Understand the Layerpath product model](/get-started/what-is-layerpath).
---
# How Brand memory affects creation
Canonical URL: /brand-memory/how-it-affects-creation
Description: What changes in the projects you create after Brand memory is set up, and why projects you already made keep the look they were created with.
When Layerpath creates a project, it captures the active Brand memory for that project. Brand memory guides the identity and wording Layerpath generates for new projects; projects that already exist keep the version captured when they were created.
Brand memory is Beta. It provides guidance to generation rather than a guarantee. Review wording, visual choices, and voice settings in the editor before you share a project.
## What Brand memory can guide [#what-brand-memory-can-guide]
The two creation paths can use the active Brand memory when they create a project. The generated parts differ by path.
| Path | What Brand memory shapes |
| ---------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Create a project with Agent](/create/create-with-agent) | Titles and body copy in the plan Layerpath drafts from your prompt, then callouts, step titles, connecting prose, and voiceover scripts as the tour is built out |
| [Record your screen](/create/record-with-chrome-extension) | callouts on each step, step titles, connecting prose, and the voiceover script written after Layerpath analyzes your recording |
Both paths also take your visual identity, so the colors, logo, and fonts are applied to the project as it is created.
## What Brand memory does not do [#what-brand-memory-does-not-do]
* It does not rewrite projects you already created.
* It does not change voiceover audio that has already been generated. Existing audio is a finished file.
* New Brand voice results do not replace the voice already in use. If no Brand voice is in use, Layerpath uses default guidance until you choose **Apply brand voice**.
* It guides generation. It does not guarantee that every generated element follows it, which is why reviewing the result in the editor still matters.
## Why projects you already made do not change [#why-projects-you-already-made-do-not-change]
When Layerpath creates a project, it records the active Brand memory and keeps that record with the project. This includes identity settings such as watermark, default language, and default voice, as well as Brand voice. The project renders from its own copy afterwards, in the editor, in the player, on a shared link, and in an exported video.
Setting up Brand memory later, or changing it, does not reach back into that copy.
So if you create two projects and then set up Brand memory:
| Project | How it looks |
| ------------------------------- | ----------------------------------------------------------- |
| The two you created first | The Brand memory captured when each was created, unchanged |
| Projects created after a change | The active Brand memory at the time each project is created |
This is deliberate. Finished work that has been shared with customers stays as it was when you shared it.
There is currently no action that applies a newer Brand memory to a project you made earlier. Create a new project if that project needs the current Brand memory.
## Check the result [#check-the-result]
Brand memory changes what Layerpath generates, so review these in the editor before sharing:
* Step titles and callout wording read the way your company writes.
* Preferred and avoided terms from **Language rules** are reflected in generated wording where appropriate.
* The voiceover script reflects the tone you intended for the project.
* Colors, logo, and fonts are the ones you set on the Brand identity tab.
* The generated wording fits the audience and purpose you intended.
When wording is wrong across several projects rather than in one place, correct the voice itself on [Review your brand voice](/brand-memory/review-your-brand-voice) instead of editing each project.
---
# Review your brand voice
Canonical URL: /brand-memory/review-your-brand-voice
Description: Read what Layerpath learned about how your company writes, correct what sounds wrong, then put the voice into use.
The Brand voice tab shows what Layerpath learned about the way your company writes, and two worked examples so you can judge it. Read the examples, correct anything that sounds wrong, then choose **Apply brand voice** to put it into use.
Brand memory is Beta. Read the generated wording before you put a voice into use, and check the results in the editor after you create your next project.
## Nothing changes until you choose to use it [#nothing-changes-until-you-choose-to-use-it]
Layerpath keeps new results separate from the settings your projects use now. New results are marked as not yet in use, and the projects you create continue to use your current settings until you choose **Apply brand voice**.
This means you can read everything on this tab, correct it, and read it again, without affecting anyone's work in the meantime.
## How to read the two examples [#how-to-read-the-two-examples]
Under the heading that asks how Layerpath will write for you sit two cards. Each shows a sample screen, a callout, and the voiceover script Layerpath would write.
The pair is deliberate. **Position your product** is how you communicate value to someone who has not bought yet. **Explain how it works** is how you guide someone already using the product. Most companies sound different in those two situations, and showing both is the quickest way to check that Layerpath understood the difference.
Choose **Play voiceover** to play a script aloud in your default AI voice.
Feedback on these cards is optional, and the two responses do not do the same thing. **Yes, sounds right** records your agreement and teaches Layerpath nothing. **Needs changes** is the response for a correction.
## Correct what sounds wrong [#correct-what-sounds-wrong]
Because both examples come from one underlying voice, a correction to either one rewrites both together.
Choose **Needs changes** on the example that reads badly. Pick a reason, or describe the problem in your own words, then choose **Review change**.
Layerpath states the change it wants to make to your voice, then shows both examples rewritten so you can compare the previous and current wording side by side.
Read the proposed change itself, not only the rewritten samples. It describes how future generated wording may change, not just these two examples.
Choose **Accept change** to keep it, **Refine** to describe what is still wrong, or **Cancel** to discard it.
Nothing changes in the results until you choose **Accept change**. The correction affects new projects only after you also choose **Apply brand voice**. A correction is never applied on your behalf.
## What Layerpath learned [#what-layerpath-learned]
Below the examples, the tab lays out its understanding of your writing as a set of cards.
| Card | What it holds |
| -------------------- | --------------------------------------------------------------------------- |
| **How you write** | Your tone, rhythm, and style, with the adjectives Layerpath matched to them |
| **Position** | How you present value to prospective customers |
| **Enable** | How you guide customers who are already using the product |
| **Writing patterns** | The recurring constructions that make your writing recognizable |
| **Language rules** | The words to prefer, avoid, and keep consistent |
Most of these cards are Layerpath's reading of your company, useful to check for accuracy. **Language rules** holds the terms Layerpath uses to guide generation. Check it when a preferred or avoided word matters.
Open **More brand details** for the company overview, positioning, messaging, writing detail, and the evidence behind each conclusion. Each card names the source it came from.
## Put the voice into use [#put-the-voice-into-use]
Choose **Apply brand voice** at the top of the tab. From that point, projects you create use this voice for callouts, step titles, connecting prose, and voiceover scripts.
Until you apply these results, new projects keep using the Brand voice already in use. If your workspace has no Brand voice in use, Layerpath uses default guidance.
## Known limitations [#known-limitations]
* Corrections are held with the new results and reach generation only after you accept them and put the voice into use.
* A voice built from too little material is not used for generation. When that happens, the tab says so and offers **Change source** or manual setup instead.
* Putting a voice into use does not change projects you already created, and does not regenerate voiceover audio that already exists.
## When the result is not what you expect [#when-the-result-is-not-what-you-expect]
### The voice does not sound like your company [#the-voice-does-not-sound-like-your-company]
Choose **Needs changes** on whichever example is furthest off and describe the problem in the free text field. The quick reasons cover tone and length. Wording that is specific to your company usually needs the field.
### You accepted an example and then noticed a problem [#you-accepted-an-example-and-then-noticed-a-problem]
Choose **Needs changes** on it anyway. Layerpath warns you that the correction rewrites both examples, then proposes the change for your review as usual.
### You corrected the voice but new projects read the same way [#you-corrected-the-voice-but-new-projects-read-the-same-way]
Check that you chose **Accept change** on the proposed change, and then **Apply brand voice**. Both are required. Reading and correcting alone does not change what Layerpath generates.
### The tab says the voice was not used for generation [#the-tab-says-the-voice-was-not-used-for-generation]
Research found too little to work from. Choose a different address if a better one exists, or set the voice up by hand.
### You want to start over [#you-want-to-start-over]
Discard the new results to go back to the settings currently in use. The control asks you to confirm before anything is removed.
---
# Set up Brand memory
Canonical URL: /brand-memory/set-up-brand-memory
Description: Set up Brand memory from a website URL, a brand PDF, or your own assets before you create your next project.
Brand memory holds two things Layerpath reuses when it writes and designs for you: an **identity**, which is your logo, colors, fonts, and voice settings, and a **brand voice**, which is the words and tone your company uses. Set it up from a URL, a PDF, or your own assets before you create the next project.
Brand memory is Beta. Review the identity and voice it produces before you rely on either one, and check generated wording and visual choices in the editor.
## Where to find it [#where-to-find-it]
Open **Brand memory** in the workspace navigation. The page has **Brand identity** and **Brand voice** tabs and opens on Brand identity.
## What you need [#what-you-need]
* A workspace where you can edit settings. Viewer access is read only.
* A company website URL, a PDF with brand material, or the assets and writing examples you want to enter yourself.
| Way to start | What it does |
| ------------------------------ | --------------------------------------------------------------------------------------------------------------- |
| Website URL, then **Generate** | Prepares your identity from the website. After you apply it, Layerpath prepares Brand voice results for review. |
| **Upload PDF** | Uses a style guide, design system, or brand-guidelines PDF up to 25 MB. |
| **Create manually** | Lets you enter identity assets and provide brand-voice examples yourself. |
## Create with a URL [#create-with-a-url]
Enter your website address and choose **Generate**. Layerpath prepares your identity for review. After you apply identity, it prepares Brand voice results for review.
When the result is ready, review the Brand identity tab and choose **Apply identity** only after the logos, colors, icons, fonts, language, voice, and watermark are correct.
## Upload a PDF [#upload-a-pdf]
Choose **Upload PDF**, then add a PDF of your style guide, design system, or brand guidelines. PDF upload is available to everyone.
The visible stages show which part of preparation is running. Choose **Stop** only when you want to end a running job. Once the Brand identity tab is ready, correct it and choose **Apply identity**.
## Create manually [#create-manually]
Choose **Create manually** when you want to upload identity assets and set the workspace details yourself.
Use **How you position your product** to provide the value Layerpath should write about. You can also add **How you explain your product**, tone characteristics, and preferred or avoided words. Choose **Generate brand voice** when you are ready to prepare the results for review.
## Review the identity [#review-the-identity]
Once the identity arrives, work through the sections on the Brand identity tab and correct anything that is wrong.
Confirm the logos, icons, and colors. Each color carries a hex value, so you can check it against your brand reference directly.
Choose **Add color** to add a color Layerpath did not find.
Layerpath sets a heading typeface and a body typeface. Both are shown with a preview of the name.
**Language** sets the language for generated voiceovers. **Voice** picks the AI voice for the workspace. Choose **Preview** to hear it before you apply the identity.
The watermark toggle shows or hides the Layerpath watermark in your projects.
## What happens next [#what-happens-next]
After you apply identity, new projects use it. Layerpath then learns Brand voice in the background and prepares results for you to review on the Brand voice tab.
Continue with [Review your brand voice](/brand-memory/review-your-brand-voice) when that notice appears.
## Choices and variations [#choices-and-variations]
| Situation | What to do |
| ------------------------------------------------------- | --------------------------------------------------------------------------------- |
| You would rather not share a website | Choose **Create manually** and supply your logo, colors, and fonts yourself. |
| Your website changed | Choose **Change source** to run research against a different address. |
| Layerpath could not reach your site | Check the address, then choose **Create manually** if the site stays unreachable. |
| The address is not your own company site | Layerpath asks you to confirm before it spends research on it. |
| Your workspace inherits branding from another workspace | The tab is read only. Switch to custom branding to edit it. |
| You have Viewer access | The tab is read only. Ask an Owner or Admin to make the change. |
## Known limitations [#known-limitations]
* Brand memory shapes projects you create from now on. It does not repaint projects that already exist. [How Brand memory affects creation](/brand-memory/how-it-affects-creation) explains why.
* The voice is not used for generation until you choose **Apply brand voice** on the Brand voice tab.
* A PDF can provide identity and writing guidance without enough product context to prepare the two Brand voice examples.
* Research reads your public website. It cannot read pages behind a sign-in.
## When the result is not what you expect [#when-the-result-is-not-what-you-expect]
### Research is still running well past 10 minutes [#research-is-still-running-well-past-10-minutes]
Leave it running and close the page if you want to. Layerpath emails you when the result is ready. If nothing arrives and the page still shows work in progress much later, note the time and the website address, then open **Help** in Layerpath.
### The colors or logos are wrong [#the-colors-or-logos-are-wrong]
Correct them on the Brand identity tab. Research reads a website, so an out of date or inconsistent site produces an inaccurate identity. Supplying the assets yourself is faster than repeated research when the site itself is the problem.
### Research finished but the Brand voice tab has nothing to read [#research-finished-but-the-brand-voice-tab-has-nothing-to-read]
The identity finishes before the voice. Wait for the notice that voice results are ready. If the voice never arrives, choose **Change source** to run research against a different address.
### You cannot edit anything on the tab [#you-cannot-edit-anything-on-the-tab]
Check whether you have Viewer access or whether the workspace inherits branding from another workspace. Both make the tab read only, and the page names the workspace it inherits from.
---
# Choose how to start a project
Canonical URL: /create/choose-how-to-start
Description: Choose Create with Agent for existing media or Record your screen for a browser workflow.
Choose **Create with Agent** when you already have images or video. Choose **Record your screen** when you need to capture a browser workflow as you perform it. Both paths create a project with one editable tour made of ordered steps.
## What you need [#what-you-need]
Decide what source you have before you start:
* Existing images, video, or reusable media → **Create with Agent**.
* A browser workflow that has not been captured yet → **Record your screen**.
Keep the outcome narrow. “Show a new customer how to connect a calendar” gives the first tour a clearer purpose than “Explain the product.”
## Start the project [#start-the-project]
On **Home**, find **Choose how to start your project**.
Select **Create with Agent** for media you already have. Select **Record your screen** for a browser workflow you need to capture in order.
**Create with Agent** opens **Start with media or a link**. **Record your screen** continues in the Chrome extension.
## What happens next [#what-happens-next]
Create with Agent asks for source media and the outcome viewers should reach. Record your screen captures the browser workflow. Layerpath then prepares an initial tour for the project.
If you are starting from media, continue with [Create a project with Agent](/create/create-with-agent).
Create with Agent
## Choose based on the source [#choose-based-on-the-source]
| What you have | Start with | Why |
| ----------------------------- | ---------------------- | ------------------------------------------------------------ |
| Product screenshots or images | **Create with Agent** | Arrange existing media around one viewer outcome. |
| An existing video | **Create with Agent** | Use the video as source material for the initial tour. |
| A Loom link | **Create with Agent** | Use **Import Loom**. |
| A live browser task | **Record your screen** | Capture the actions in the order viewers should follow them. |
* **Record your screen** captures browser workflows through the Chrome extension.
* Use **Import Loom** for a Loom link.
* A preview can appear before processing is complete. Check the visible processing state before editing.
## When the result is not what you expect [#when-the-result-is-not-what-you-expect]
### You opened the wrong path [#you-opened-the-wrong-path]
Select **Back**, return to **Choose how to start your project**, and choose the other path.
### A starting choice is missing [#a-starting-choice-is-missing]
Return to **Home** and check the available creation actions. If the action still does not appear, open **Help** from the Layerpath navigation.
### Recording does not begin [#recording-does-not-begin]
Return to Layerpath and open **Help**. Include what happened after you selected **Record your screen** and any message shown by the extension.
The starting path changes how Layerpath receives source material, not the object it creates. The project contains one editable tour, and that tour contains ordered steps. A guide mirrors the tour, while a video export renders it.
Read [What is Layerpath?](/get-started/what-is-layerpath) for the complete product model or [Understand project processing](/create/understand-processing) before reviewing an early preview.
---
# Create a project with Agent
Canonical URL: /create/create-with-agent
Description: Combine source material with a creation prompt to prepare an editable initial tour.
Create with Agent requires both source material and a creation prompt. The source shows the workflow. The prompt states what the tour should help viewers understand or accomplish.
## What you need [#what-you-need]
Prepare:
1. Source material that shows the useful workflow in its intended order.
2. A creation prompt naming the audience, task, and expected result.
3. A final success or confirmation state when the workflow has one.
Remove repeated, irrelevant, outdated, or sensitive material before adding the source.
## What Layerpath can use during creation [#what-layerpath-can-use-during-creation]
* **Visible source content and sequence** help establish what happens and in what order.
* **Useful narration in a source video** can provide additional context about intent, terminology, sequence, and important moments. It does not guarantee a better result.
* **The creation prompt** explains what the tour should teach and where it should finish.
* **Steps** controls granularity for supported video creation. **Voiceover** and **Tone** refine the requested presentation when you change them from Auto.
* **Brand memory Beta**, when configured, can provide bounded workspace identity and voice guidance.
Layerpath may not apply every input exactly as you expect. Compare the initial tour with the source and creation prompt.
## Create the project [#create-the-project]
## What usually produces a stronger initial tour [#what-usually-produces-a-stronger-initial-tour]
* Include the complete workflow, including its final success state.
* Keep images in the order viewers should follow when possible.
* Keep useful spoken narration in a source video.
* State one narrow viewer outcome.
* Remove repeated screens, long pauses, unrelated states, and sensitive information.
These choices are guidance, not enforced requirements. A clear source reduces ambiguity, but the resulting tour still needs review.
## Choose the source [#choose-the-source]
Uploaded and library images and videos can be combined in one creation. A Loom or YouTube URL cannot currently be combined with other media. Use that URL as the only source.
| Choice | Use it when |
| ----------------------- | ---------------------------------------------------------------------------------------------------------- |
| **Upload from device** | The images or video are stored on your device. |
| **Choose from library** | The source media already exists in the Layerpath media library. Provide the current creation prompt again. |
| **Import Loom** | Your source is a Loom link. |
Library reuse provides convenient access to saved media. Do not assume that a library item preserves or reuses interaction details from an earlier Chrome extension recording. Provide the current creation prompt and review the result as a new creation.
## Decide whether to change advanced options [#decide-whether-to-change-advanced-options]
| Choice | Keep or select it when |
| ------------------------------ | --------------------------------------------------------------------------------- |
| **Steps: Auto** | Layerpath should choose the initial level of detail for supported video creation. |
| **Steps: Short** | A supported video workflow needs a more compact starting sequence. |
| **Steps: Detailed** | A supported video workflow needs more intermediate moments represented. |
| **AI voiceover: Auto** | You want the default creation behavior. |
| **AI voiceover: On** | The initial tour should include generated narration. |
| **AI voiceover: Off** | The initial tour should remain silent. |
| **Tone: Auto** | Layerpath should infer the writing style from the source and creation prompt. |
| **Clear, Friendly, or Formal** | The audience or use case requires that specific writing style. |
Advanced options refine the request. They do not replace a clear creation prompt or complete source.
When the creation contains images only, **Steps** remains on **Auto**. **Short** and **Detailed** apply to supported video creation.
## How Brand memory relates to creation [#how-brand-memory-relates-to-creation]
Brand memory is Beta. When it is configured for the workspace, Layerpath can use bounded identity and voice guidance while preparing the initial tour. This can include approved terminology, writing guidance, and identity defaults. Application can vary, so review wording, visual choices, and voice settings in the editor.
This page does not define Brand memory setup or promise that every generated element will follow it.
## What happens next [#what-happens-next]
Open the initial tour in the editor. Compare it with the source and creation prompt. Review and refine the generated steps, Intro, Outro, callouts, and requested voiceover. Verify that each is accurate before sharing the tour.
Understand processing
* Use **Import Loom** for a Loom link.
* Use a Loom or YouTube URL by itself. It cannot currently be combined with uploaded or library media.
* Follow the file guidance shown in the current upload screen.
* The preview can change until processing reaches **Done**.
* Narration, Advanced options, and Brand memory provide context. They do not guarantee a specific result.
* Verify generated steps, Intro, Outro, callouts, and voiceover before sharing the tour.
## When the result is not what you expect [#when-the-result-is-not-what-you-expect]
### The arrow button is unavailable [#the-arrow-button-is-unavailable]
Confirm that source media appears in the composer and that the creation prompt is not empty. If an upload is in progress, wait for it to finish.
### Add source is unavailable [#add-source-is-unavailable]
Select a media item in the library first. The selected-item count updates before **Add source** becomes available.
### A Loom link is not accepted [#a-loom-link-is-not-accepted]
Use **Upload from device** or **Choose from library** when either path fits your source. If you need help with the link, open **Help** from the Layerpath navigation.
### Processing does not finish [#processing-does-not-finish]
Keep the existing project and open **Help**. Include the project name, visible phase, and exact message shown.
---
# Record your screen with the Chrome extension
Canonical URL: /create/record-with-chrome-extension
Description: Capture a browser workflow as ordered steps, then review the resulting tour in the editor.
Use **Record your screen** when the browser workflow itself is the source. Layerpath turns accepted clicks into ordered steps inside one editable tour. Use [Create with Agent](/create/create-with-agent) instead when you already have images, video, a Loom or YouTube link, or library media.
## What the extension captures [#what-the-extension-captures]
Each accepted click is the main step boundary. The extension captures the visible page state for that click, the click location, and useful page context such as the page title, URL, and visible control labels.
Typing, focus changes, and scrolling can provide supporting interaction signals. They do not create steps by themselves. The typing signal does not include the characters you type, but the visible page screenshot can still contain text, form values, names, or other sensitive information.
Optional microphone narration can provide additional context while Layerpath prepares the tour. It requires microphone permission. When narration is on, audio playing in the recorded tab may also be captured. Narration can help communicate intent, terminology, sequence, and important moments, but you must review all generated content.
## What is not captured reliably [#what-is-not-captured-reliably]
* Hover-only actions do not create a step.
* Typing or scrolling without a click does not create a step.
* Drag interactions, canvas controls, and short-lived visual states may not produce a useful step.
* Dropdowns and modals are useful only when the intended state is visible at the captured click.
* Content inside embedded frames may be incomplete, especially when the embedded content comes from another site.
* Chrome settings, the Chrome Web Store, extension pages, and other restricted browser pages are not recordable.
* Other browsers are outside the current supported boundary for this page.
You can switch between normal Chrome tabs during a recording. Page access and visible state can vary between tabs, so verify every step before sharing.
## Before you start [#before-you-start]
* [Choose Record your screen as the starting path](/create/choose-how-to-start).
* Install the official [Layerpath Chrome extension](https://chromewebstore.google.com/detail/layerpath-create-interact/bckkboajpcaejmbajljhebicnflhdajc).
* Sign in to Layerpath in Chrome and confirm the workspace where the project should appear.
* Open a normal website. Chrome internal pages cannot be recorded.
* Use test data where possible. Remove passwords, personal information, API keys, private messages, and unrelated tabs from the workflow.
* Rehearse one clear path from start to finish. Deliberate actions produce a tour that is easier to review.
* Keep Chrome open until the Layerpath processing view appears after you finish.
## Record the workflow [#record-the-workflow]
On Layerpath Home, choose **Record your screen**. Go to the normal web page where the workflow starts, then open the Layerpath extension from the Chrome toolbar.
If the extension asks you to sign in, complete sign-in in the opened tab. Return to the starting page and reopen the extension.
Turn on narration only if you want to speak during the workflow. If Chrome asks for microphone permission, allow it or turn narration off before continuing.
Choose **Start recording**. On a first recording, Layerpath may show **How recording works**. Review it and choose **Start recording** again.
Wait for the countdown to finish before interacting with the page.
Select only the controls a viewer needs to follow. Each accepted click increases the step count. Use the count as an orientation aid, then verify every step in the editor after finishing.
If you need a break, open the extension and choose **Pause**. Pause stops step capture and pauses the screen and narration recording. Choose **Resume** when you are ready to continue the same recording.
Choose **Finish** when the workflow is complete. The control changes to **Finishing...** while the extension submits the recording and opens Layerpath.
The Layerpath processing view is the handoff acknowledgment. The **Finishing...** label alone does not prove that the recording reached Layerpath.
After the processing view opens, the source page is no longer needed for the active recording. Keep the Layerpath tab open to monitor processing.
Choose **Discard** only when you want to abandon the capture. At **Discard recording?**, choose **Discard** to delete the active recording or **Cancel** to return to it.
Do not choose **Discard** if you may need the capture. The current flow does not offer a restore action after you confirm deletion.
Layerpath creates a project with one editable tour. The recorded clicks become ordered steps. A preview can appear before callouts, content, and optional voiceover are complete.
Follow the visible processing stages. Treat the handoff as complete only when processing reaches **Done** and **Edit** is available.
Compare the step count and sequence with the workflow you performed. Open several steps and confirm that each one shows the expected page state and target. Then continue with [Editor basics](/edit/editor-basics).
## Review the result before sharing [#review-the-result-before-sharing]
The expected result is a new project whose tour contains the browser actions in order. Verify all of the following before sharing:
* The project appears in the intended workspace.
* The tour opens in the editor.
* The number and order of steps match the recorded path.
* Each captured step shows the expected page state.
* Hotspots and callouts point to the intended action.
* Intro, Outro, callouts, and optional voiceover are accurate and supported by the recorded workflow or your narration.
* No password, token, personal information, or unrelated browser content is visible.
* The complete tour works in Preview.
If Layerpath is still preparing the project, use [Understand processing](/create/understand-processing) to identify the visible state. Do not assume that a completed extension animation means processing has finished.
## Known limitations and common mistakes [#known-limitations-and-common-mistakes]
| Situation | What to do |
| ----------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| The capture is shorter than three seconds or contains no accepted click | Start again and record one complete, deliberate path with at least one click. A very short or empty capture may not create a project. |
| You clicked the wrong control | If the rest of the recording is useful, finish and remove the unwanted step in [Manage steps](/edit/manage-steps). Otherwise, discard and start again. |
| A dropdown, modal, or loading state disappeared too quickly | Pause before the action, prepare the state, resume, and click only after the intended state is visible. Confirm the result in the editor. |
| You need a hover, drag, canvas action, or embedded control | Record the closest reliable click state, then inspect the result. Do not assume the interaction itself was captured. |
| The page reloads during recording | Reopen the extension and check whether the recording and step count are still active. Continue only after the active state is visible, then verify the surrounding steps in the editor. |
| You close the original tab before finishing | Reopen the extension immediately. Continue only if the active recording and step count are visible. Screen and narration continuity may stop when the original tab closes. |
| You see a microphone permission error | Allow microphone access in Chrome, or turn narration off and start without it. |
## Sensitive data and privacy [#sensitive-data-and-privacy]
The extension can capture website content and user activity that is visible during the workflow. Treat every captured page as publishable source material.
* Use a test account and synthetic records whenever possible.
* Close private messages, unrelated tabs, notifications, and password managers before starting.
* Mute unrelated tab audio, calls, and videos before turning narration on.
* Do not enter passwords, access tokens, payment details, health information, or customer data during the recording.
* Remember that typed characters are not required for the page screenshot to reveal the final visible value.
* Review every step at full size before sharing the tour.
* If sensitive information appears, remove the affected step or record the workflow again with safe data.
## When the result is not what you expect [#when-the-result-is-not-what-you-expect]
### The extension shows **Restricted Page** [#the-extension-shows-restricted-page]
Open a normal website in Chrome, then reopen the extension. Chrome internal pages and extension pages are not recordable.
### The extension asks you to sign in [#the-extension-asks-you-to-sign-in]
Complete sign-in, return to the website, and reopen the extension. Confirm the intended workspace before starting.
### **Start recording** does not begin a countdown [#start-recording-does-not-begin-a-countdown]
Keep the source tab open and note the visible error. Reopen the extension once. If the countdown still does not begin, open **Help** in Layerpath and include the time, Chrome version, website domain, and exact message.
### The processing view does not appear after **Finish** [#the-processing-view-does-not-appear-after-finish]
Keep Chrome open and connected. Check **Projects** once for the new project. If it is not there, preserve the exact visible message and open **Help** in Layerpath. The extension attempts background retries, but a user-controlled retry is not currently available in the visible flow.
### Processing does not reach **Done** [#processing-does-not-reach-done]
Keep the project and note the visible stage. Read [Understand project processing](/create/understand-processing). If the stage remains unchanged, open **Help** and include the project name, time, and visible message.
### The project appears with missing or incorrect steps [#the-project-appears-with-missing-or-incorrect-steps]
Do not share it. Remove an unwanted step when the source is otherwise correct. If steps are missing or the captured page state is wrong, preserve the project and open **Help** with the affected step numbers.
---
# Understand project processing
Canonical URL: /create/understand-processing
Description: Distinguish an early tour preview from a completed project and interpret the visible processing phases.
A tour preview can appear before project creation is complete. Judge progress by the highlighted phase, not by the appearance of the preview. The initial tour is ready to inspect when the sequence reaches **Done** and **Edit** is available.
## How to know when the project is ready [#how-to-know-when-the-project-is-ready]
Check two visible signals:
1. The phase sequence reaches **Done**.
2. **Edit** is available.
Until both signals appear, callouts, requested voiceover, or final project preparation may still be in progress.
## Review progress [#review-progress]
Use the active label in the progress indicator. Do not infer readiness from how complete the preview looks.
Check whether the source media and step sequence reflect the intended workflow. During **Callouts**, expect the guidance copy to change.
Continue reviewing until the sequence reaches **Done** and **Edit** becomes available.
Select **Edit** and verify the step order, callouts, source media, and any requested voiceover.
## What happens next [#what-happens-next]
The editor opens the tour that Layerpath prepared. Review the complete sequence before changing styles or sharing it.
Learn editor basics
| Phase | What it means for your review |
| -------------- | -------------------------------------------------------------------------- |
| **Callouts** | The preview can appear while Layerpath writes and refines callout copy. |
| **Voiceover** | Voiceover work is in progress when it is part of the creation flow. |
| **Finalizing** | Layerpath shows **Getting your project ready**. Keep waiting for **Done**. |
| **Done** | Confirm that **Edit** is available, then inspect the tour. |
* Processing time can vary; this page does not provide a fixed duration.
* A visible preview is not a completion signal.
* **Edit** and **Share** can remain unavailable while a processing phase is active.
* Check requested voiceover and callout wording after **Done**.
## When the result is not what you expect [#when-the-result-is-not-what-you-expect]
### **Edit** and **Share** are unavailable [#edit-and-share-are-unavailable]
Check the highlighted phase. If **Callouts**, **Voiceover**, or **Finalizing** is active, wait for **Done**.
### Callout wording changes while you review [#callout-wording-changes-while-you-review]
The **Callouts** phase indicates that the copy is still being refined. Review it again after **Done**.
### The phase stops changing or an error appears [#the-phase-stops-changing-or-an-error-appears]
Keep the existing project and open **Help**. Provide the project name, active phase, exact message, and whether a preview is visible.
### The result looks incomplete after **Done** [#the-result-looks-incomplete-after-done]
Open **Edit** and compare the steps with the source workflow. Note which step, callout, source item, or requested voiceover is missing before opening **Help**.
---
# Editor basics
Canonical URL: /edit/editor-basics
Description: Orient yourself in the tour editor: select a step, use its contextual controls, make a change, and Preview.
Use this page to learn the manual editing loop, not every editor control.
## Choose where to work [#choose-where-to-work]
| When you need to | Start here |
| ---------------------------------------- | ------------------------------------------------------------------------ |
| Change content on the selected step | The canvas and contextual controls. |
| Change the tour order or structure | [Manage steps](/edit/manage-steps). |
| Change the step composition | [Layouts and backgrounds](/edit/layouts-and-backgrounds). |
| Change how viewers move through the tour | [Tour playback settings](/edit/tour-playback-settings). |
| Tell viewers where to act | [Hotspots, areas, and callouts](/interactivity/hotspots-areas-callouts). |
## What you need [#what-you-need]
* A project that opens in the editor. If the project is still being prepared, [check its processing state](/create/understand-processing).
* One specific change you want to make.
* Any replacement image or copy needed for that change.
## The basic editing loop [#the-basic-editing-loop]
Choose the step counter in the lower-left corner to open the step list, then select the step you want to edit. The canvas and toolbar update for the selected step.
Confirm the step type above the canvas, then read the controls below it. An Image step can expose **Insert**, **Layout**, **Voiceover**, and more actions.
If the expected control is missing, check the selected step and its type before continuing.
Choose the toolbar control that matches the change, then confirm the result on the canvas. For example, a Blank step can expose its color choices.
Choose **Preview** in the editor header when a change affects what viewers see or do. Before sharing, move through the complete tour from the beginning and check the step order, content, guidance, and ending.
Return to the editor to correct anything that does not match your intended result, then preview again.
This orientation deliberately stops after one manual change and Preview. Continue with [Manage steps](/edit/manage-steps), [Layouts and backgrounds](/edit/layouts-and-backgrounds), or [Hotspots, areas, and callouts](/interactivity/hotspots-areas-callouts) for a focused walkthrough.
* The toolbar is contextual, so controls change with the selected step and element.
* This page covers manual orientation only; it is not a complete editor reference.
* **Agent** opens Edit with Agent, a separate Beta editing path that this page does not teach.
* Preview the complete tour before sharing.
## When the result is not what you expect [#when-the-result-is-not-what-you-expect]
### The toolbar does not show the control you expected [#the-toolbar-does-not-show-the-control-you-expected]
Confirm the selected step and its type. The toolbar is contextual. Return to the step list, select the intended step, and check the toolbar again.
### You cannot find the step you need [#you-cannot-find-the-step-you-need]
Open the step selector again and confirm the step label above the canvas. If the step is still missing, stop editing and contact support with the project and expected step.
### A change is missing in **Preview** [#a-change-is-missing-in-preview]
Return to the editor and reopen the affected step. Confirm that the editor shows the intended value, then preview again. If the editor and Preview still differ, stop and contact support with the project, step, and visible mismatch.
## What to do next [#what-to-do-next]
Manage steps
---
# Layouts and backgrounds
Canonical URL: /edit/layouts-and-backgrounds
Description: Place step media and choose a supporting color or abstract background.
Choose a composition that keeps the step media readable and leaves enough room for its guidance.
## What you need [#what-you-need]
* A step where **Layout** is available.
* Media that is already cropped closely enough to remain useful at the chosen size.
* A clear reason to leave space beside or above the media.
## Choose the composition [#choose-the-composition]
Select the step, then choose **Layout** in the contextual toolbar. If the control is missing, check the step type before continuing.
Choose one of the five visible positions:
| Layout | Use it for |
| ---------- | ----------------------------------------------------- |
| **Full** | Media that should occupy the available canvas. |
| **Right** | Media on the right with supporting space on the left. |
| **Left** | Media on the left with supporting space on the right. |
| **Bottom** | Media below supporting space. |
| **Top** | Media above supporting space. |
If a fit control is available, choose **Cover** to fill the media area or **Fit** to keep the complete media visible. Check the canvas for unwanted cropping or unused space.
Choose **Colors** for a quiet field or **Abstract** for one of the available image backgrounds. Keep the media, guidance, and any text readable against the result.
Preview the previous step, the edited step, and the next step. Confirm that the composition does not create an abrupt scale or contrast change.
* Layout availability can change with the selected step type.
* Intro and Outro have additional theme behavior that this page does not cover.
* Always use **Preview** to check the final crop, spacing, and guidance position.
## When the composition is not what you expect [#when-the-composition-is-not-what-you-expect]
### Important content is cropped [#important-content-is-cropped]
Switch from **Cover** to **Fit**, choose a different layout, or replace the source with a closer crop.
### The background competes with the step [#the-background-competes-with-the-step]
Return to **Colors** and choose a quieter background. The viewer action should remain the strongest visual element.
### Layout is missing [#layout-is-missing]
Confirm the selected step and its type. If **Layout** is still unavailable, choose another step where **Layout** is available or contact support with the project and step.
## What to do next [#what-to-do-next]
Configure tour playback
---
# Manage steps
Canonical URL: /edit/manage-steps
Description: Add, reorder, duplicate, hide, and delete the ordered steps in a tour.
Build a clear tour sequence by adding only the steps viewers need and placing them in the right order.
## What you need [#what-you-need]
* A project that opens in the editor.
* A simple outline of the viewer journey from first step to last step.
* Any media needed for a new Image step or Video step.
## Arrange the tour [#arrange-the-tour]
Choose the step counter in the lower-left corner. Select a step to inspect its content and its place in the sequence.
At the intended insertion point, choose **Add step** to add sourced media or **Blank step** to begin with an empty canvas. Confirm the new step appears in the step list.
Drag a step thumbnail to its intended position. Read the full sequence from top to bottom, then choose **Preview** and move through it in the same order a viewer will.
Open the menu for the selected step, then choose the action that matches your intent:
| Action | Use it when |
| ------------------ | ---------------------------------------------------------------------------------------- |
| **Duplicate step** | You want a copy to adapt without rebuilding the step. |
| **Hide step** | You want to retain the step in the editor while you test the sequence without it. |
| **Delete step** | You no longer need the step. Confirm the deletion only after checking the selected step. |
Start at the beginning. Check that every step has a clear purpose, the order is logical, and the tour reaches the intended result without unrelated content.
* Hidden step behavior should be checked in **Preview** before the tour is shared.
* Deletion is permanent after confirmation. Duplicate first when you want a safe copy to compare.
* This page does not cover bulk selection or step-specific media editing.
## When the sequence is not what you expect [#when-the-sequence-is-not-what-you-expect]
### A new step appears in the wrong place [#a-new-step-appears-in-the-wrong-place]
Open the step list, drag the step to the intended position, and preview again.
### A hidden step still affects the experience [#a-hidden-step-still-affects-the-experience]
Return to the step list and confirm that you hid the intended step. Preview from the beginning. If the step still appears, contact support with the project and step name.
### You deleted the wrong step [#you-deleted-the-wrong-step]
Stop editing and contact support. Do not make additional structural changes until you know whether the step can be recovered.
## What to do next [#what-to-do-next]
Add viewer guidance
---
# Tour playback settings
Canonical URL: /edit/tour-playback-settings
Description: Configure navigation, auto zoom, motion, autoplay, cursor, and callout visibility.
Set how viewers move through the tour, then preview the complete experience before sharing it.
## What changes what? [#what-changes-what]
| Setting | What it changes |
| ----------------------- | ------------------------------------------------------------------- |
| **Show navigation bar** | Whether viewers can see progress and move directly to another step. |
| **Auto zoom** | Automatic framing, when the control is available. |
| **Motion speed** | How quickly framing moves between steps. |
| **Enable autoplay** | Whether Layerpath advances the tour after the viewer begins. |
| **Cursor** | The cursor style shown to viewers. |
| **Callouts** | Visibility of callouts, arrows, and Back and Next buttons. |
## What you need [#what-you-need]
* A tour with its final step order.
* A decision about whether viewers should control the pace or follow an automatic sequence.
* The screen sizes your audience is likely to use for Preview review.
## Configure playback [#configure-playback]
Open the selected step menu, choose **Settings**, then open **General**.
Turn on **Show navigation bar** when viewers should see progress and jump directly to another step.
If **Auto zoom** is available, turn it on to let Layerpath frame steps automatically. Choose **Slow**, **Smooth**, or **Quick** for the motion between steps.
Auto zoom is a tour setting. An [authored zoom](/interactivity/hotspots-areas-callouts#add-an-authored-zoom) is a focus region you add to one step.
Turn on **Enable autoplay** when the tour should control the pace after the viewer begins. Choose 3, 5, 7, or 9 seconds per step.
Silent image steps advance after the selected duration. A step with voiceover waits for the voiceover to finish, even when it runs longer than the selected duration.
Choose **Default**, the pointer, or the circle. Use the style that makes the intended action easy to follow without covering the target.
Open **Callouts** in Tour Settings. Choose whether to show callouts, arrows, and the Back and Next navigation buttons.
Move through the complete tour. Check direct navigation, motion, step timing, cursor placement, callout readability, and Back and Next behavior.
## How autoplay works [#how-autoplay-works]
* A viewer begins the tour by interacting with the first screen. Autoplay does not start before that first action.
* Silent image steps use the duration selected in **Tour Settings**.
* Image steps with voiceover advance after the voiceover finishes.
* The tour stops after the final step. It does not automatically restart.
- Browsers can block audio until the viewer interacts with the page. If audio cannot play, the tour can continue using the step duration.
- Auto zoom may not be available in every environment.
- Settings shown here describe the tour. This page does not establish behavior for other outputs.
- Preview the complete tour at the screen sizes your audience is likely to use.
## When playback is not what you expect [#when-playback-is-not-what-you-expect]
### The tour advances too quickly [#the-tour-advances-too-quickly]
Increase the duration per step or turn off autoplay. Preview steps with longer voiceover before choosing a duration.
### The share link does not begin automatically [#the-share-link-does-not-begin-automatically]
Interact with the first screen to begin playback. Browsers require this first action before Layerpath can start voiceover reliably.
If the tour still does not advance, return to **Tour Settings** and confirm **Enable autoplay** remains on. Preview the tour again before reopening the share link.
### The tour waits on one step [#the-tour-waits-on-one-step]
Turn off autoplay and Preview the tour again. If the tour still waits unexpectedly, contact support with the project and affected step.
### Framing moves too quickly [#framing-moves-too-quickly]
Choose **Slow** or **Smooth**. If one step still needs a specific focus region, add an authored zoom to that step.
### Navigation or callout controls are missing [#navigation-or-callout-controls-are-missing]
Return to Tour Settings and check both **General** and **Callouts**. Preview again after changing the relevant toggle.
## What to do next [#what-to-do-next]
Refine viewer guidance
---
# Create your first project
Canonical URL: /get-started/create-your-first-project
Description: Choose Create with Agent or Record your screen, then review the resulting tour in the editor.
## Choose your starting point [#choose-your-starting-point]
Layerpath has two entry points. Start from material you already have, or record the browser workflow as you perform it.
### Start with media or a link [#start-with-media-or-a-link]
Use **Create with Agent** when the source material already exists. Choose one source path:
* **Upload from device** for new images or video.
* **Choose from library** to reuse media already saved in your workspace. Choose only the items needed for the viewer outcome.
* **Import Loom** for a Loom link.
Add the source, then describe what the project should help viewers do.
[Create a project with Agent](/create/create-with-agent).
### Record a workflow [#record-a-workflow]
Choose **Record your screen** when the browser workflow is the source you need to capture. Perform the task in its intended order and keep unrelated tabs, personal data, and credentials out of the recording.
After the recording is added to Layerpath, review the resulting tour in the editor.
## Create the project [#create-the-project]
Write one sentence that names the audience, the task, and the intended result. Keep the first project narrow enough to review from start to finish.
Add new media or a Loom or YouTube link, choose existing workspace media, or record the browser workflow. Include only what the viewer needs.
Open the project in the editor. Check that the steps are accurate, useful, and in the intended order.
Experience the tour from the beginning. Confirm that each step makes the next action clear before sharing it with viewers.
A preview can appear while a project is still being prepared. Check the visible processing state before treating the initial tour as complete.
Check the step order, accuracy, sensitive information, and next action before sharing it as an interactive tour.
1. Confirm that the tour opens with the information viewers need and ends with the intended result.
2. Remove unrelated steps and put the remaining steps in a useful order.
3. Check every step for private information or outdated source media.
4. Make sure the guidance explains the next action without repeating what is already clear on screen.
5. Preview the complete tour as a viewer would experience it.
## What happens next? [#what-happens-next]
* If the project is still being prepared, read [Understand project processing](/create/understand-processing).
* To review the canvas, step list, and preview controls, continue to [Editor basics](/edit/editor-basics).
* To compare interactive tour, guide, and video export, read [What is Layerpath?](/get-started/what-is-layerpath#choose-how-people-experience-it).
Create with Agent
---
# What is Layerpath?
Canonical URL: /get-started/what-is-layerpath
Description: Turn software knowledge into reusable guidance for customer education and internal training. Learn how projects, tours, guides, and video exports relate.
Layerpath turns software knowledge into reusable guidance for customer education, onboarding, support, and internal software training. Update the steps that changed to keep the guidance current as your software changes.
Each workflow becomes a **project** that people can experience as an **interactive tour**, a **guide**, or a **video export**. You edit one tour made of ordered **steps**; each output uses that source.
## Three ideas to remember [#three-ideas-to-remember]
1. A **project** is the item you create and organize in a workspace.
2. The **tour** is the source you edit inside that project.
3. A **step** is one ordered part of the tour.
Edit the tour when you need to change the content, sequence, or guidance. The other formats use that same source.
## Choose how people experience it [#choose-how-people-experience-it]
### Interactive tour [#interactive-tour]
Use an interactive tour when viewers should move through the workflow one step at a time.
### Guide [#guide]
Use a guide when viewers should scan or read the steps as a document. The guide mirrors the tour; it does not have a separate editor.
### Video export [#video-export]
Use a video export when viewers should watch a rendered version without interacting with the tour.
A project has one editable tour. A guide mirrors that tour, and a video export renders it. Change the tour when the content needs to change.
| Term | What it means |
| ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------- |
| **Workspace** | The container for projects and workspace-level areas such as team folders, Playlists, Brand memory (Beta), Insights, members, and settings. |
| **Project** | The top-level item you create, find, and organize. Each project contains one editable tour. |
| **Tour** | The source you edit inside a project. It contains the ordered steps. |
| **Step** | One ordered part of a tour. The step types are Intro, Image step, Video step, Blank step, and Outro. |
| **Guide** | A reading presentation that mirrors the tour and its steps. |
| **Video export** | A rendered version of the tour. |
A guide can also be copied as steps or downloaded as a PDF. A hotspot, Area, or callout can add guidance within a step; these are not additional step types.
## One project, three viewer experiences [#one-project-three-viewer-experiences]
For example, one customer-onboarding project can keep a single tour as its source. The same ordered steps can support an interactive tour, a guide for a help center, and a video export for asynchronous training.
Create a project
---
# Use Layerpath in ChatGPT
Canonical URL: /integrations/chatgpt
Description: Create interactive tours from your screens in ChatGPT, then edit, share, or export them as videos without leaving the chat.
Layerpath turns real software screens into interactive tours with hotspots and callouts. In ChatGPT, attach screenshots or a screen recording, or paste a Loom or YouTube link. Ask Layerpath to create a tour, then edit it, add voiceover, share it, or export it as a video in ChatGPT.
## What you need [#what-you-need]
* A ChatGPT account.
* A Layerpath account. You can create one when you sign in.
* Owner, Admin, or Editor access to a Layerpath workspace. Viewers can't create tours.
## Add Layerpath to ChatGPT [#add-layerpath-to-chatgpt]
1. In ChatGPT, open **Plugins** in the sidebar and search for Layerpath. Select **Install plugin**. If ChatGPT asks, select **Connect**.
2. Sign in to Layerpath with Google or your email and password, or create an account.
3. Select the workspace where tours created in ChatGPT will be saved, then select **Allow**. You'll return to ChatGPT.
Ask for a tour in any ChatGPT chat. You can also type **@** and select Layerpath.
## Make a tour [#make-a-tour]
Attach screenshots or a screen recording, or paste a Loom or YouTube link into the chat. Then ask for a tour. For example:
* Make an interactive tour from these screenshots
* Turn this Loom recording into a step-by-step tour
ChatGPT may ask you to confirm before Layerpath creates a tour. If your message includes attached screens or a link, Layerpath starts creating the tour after any required confirmation.
If no screens are attached, Layerpath shows a **Create with Agent** box in the chat. Drop screenshots or a screen recording on **Drop files here**, or select it to choose files. You can also paste a Loom or YouTube link in the box. The tour starts when the files are added. If it doesn't, select the arrow to start it.
Layerpath shows creation progress, then the tour's steps. Creating a tour usually takes a few minutes; longer videos may take longer. Asking again creates a separate tour.
**Where Layerpath appears**
On supported ChatGPT plans, select **Open editor** to open Layerpath beside the chat. Select a step to edit it or choose **Play** to preview the tour. In **Create with Agent**, drop screenshots or a screen recording on **Drop files here**, or select it to choose files for another tour. You can also select **Edit** on a tour. On ChatGPT Free and Go, Layerpath appears in the chat only.
## What you can ask [#what-you-can-ask]
* Create a tour from screenshots, a screen recording, or a Loom or YouTube link.
* Shorten steps, clarify instructions, or change their tone.
* Translate a tour in place or create a translated copy.
* Add voiceover in a voice or language you choose.
* Blur private information, such as names or email addresses.
* Move, merge, hide, or delete steps. Change the intro, outro, or tour name.
* Zoom in on part of a step.
* Duplicate a tour to customize it for a customer.
* Undo a change.
* Find an earlier tour by name and open it.
* Get a tour's share link or embed link.
* Export a tour as a video.
For some changes, Layerpath shows proposed edits first. Select **Go ahead** to apply them, or reply in the chat to revise the plan.
Layerpath needs screens from existing software. It doesn't design screens for software that hasn't been built.
## Where your tours are saved [#where-your-tours-are-saved]
Tours created in ChatGPT are saved in the Layerpath workspace you selected during connection. They appear under your account alongside your other tours at app.layerpath.com, where workspace members can open them.
Layerpath publishes new tours at a share link when they're ready. Anyone with the link can open them. For an unpublished tour, asking ChatGPT for a share link or selecting **Copy link** publishes it.
To switch workspaces, disconnect Layerpath in ChatGPT, reconnect it, and select the other workspace.
## Who can change a tour [#who-can-change-a-tour]
You can edit tours you created. Owners and Admins can edit any tour in the workspace. **Undo** reverses changes made through ChatGPT, not edits made directly in Layerpath.
## Disconnect Layerpath [#disconnect-layerpath]
**In ChatGPT:** open **Settings**, select **Plugins**, choose Layerpath, then select **Disconnect**.
**In Layerpath:** Owners and Admins can open **Settings**, then **Integrations**. The ChatGPT card lists connected members, each with a **Disconnect** option. Once disconnected, that person must reconnect to use Layerpath in ChatGPT.
Tours already created remain in the workspace.
## If something goes wrong [#if-something-goes-wrong]
* **"Your Layerpath session has expired":** reconnect Layerpath in ChatGPT.
* **"You're no longer a member of this Layerpath workspace":** reconnect Layerpath and select a workspace you belong to.
* **"This tour belongs to another of your Layerpath workspaces":** reconnect Layerpath and select the workspace containing that tour.
* **"Only the tour's creator or an Owner or Admin in (your workspace) can edit it":** ask the tour's creator or a workspace Owner or Admin to make the change.
* **You can't select a workspace when connecting:** if you're a Viewer, ask an Owner or Admin to change your role to Editor.
* **"These files couldn't be used to create a tour":** attach screenshots or a screen recording, or paste a Loom or YouTube link.
* **A share link doesn't open:** the tour may be unpublished. Ask ChatGPT for its share link to publish it.
To reconnect, open **Settings**, select **Plugins**, choose Layerpath, then select **Reconnect**. If you don't see it, disconnect and connect again.
## Limits [#limits]
* Layerpath accepts up to 30 files from the chat at a time. Each file can be up to 200 MB.
* In the **Create with Agent** box, screenshots can be up to 15 MB each. Screen recordings can be up to 200 MB and 10 minutes long.
* Loom and YouTube videos can be up to 10 minutes long.
* AI can make mistakes. Review your tour before sharing it.
## Privacy [#privacy]
Layerpath receives each request as ChatGPT words it, including screenshots or recordings it passes on and the tour it's about. ChatGPT may include details from your chat or its memory. Layerpath can't read the rest of your conversation or your ChatGPT memory. ChatGPT also sends your language, time zone, and approximate location, which Layerpath doesn't use or keep. Screens added to the **Create with Agent** box go directly to Layerpath. Layerpath doesn't use any of this to train AI models.
Read how Layerpath handles your data: [https://www.layerpath.com/privacy](https://www.layerpath.com/privacy)
## Get help [#get-help]
Email [support@layerpath.com](mailto:support@layerpath.com). Include that you were using ChatGPT, the tour name or share link, your request, what happened, and the time with your time zone.
---
# Use Layerpath in Claude
Canonical URL: /integrations/claude
Description: Create interactive tours from your screens in Claude, then edit, share, or export them as videos without leaving the chat.
Layerpath turns real software screens into interactive tours with hotspots and callouts. In Claude, add screenshots or a screen recording, or share a Loom link. Layerpath shows the tour in the chat, where you can edit it, share it, or export it as a video.
## What you need [#what-you-need]
* A Claude account. Claude Free supports one custom connector.
* A Layerpath account. You can create one during connection.
* Owner, Admin, or Editor access to a Layerpath workspace. Viewers can't create tours.
## Connect Layerpath [#connect-layerpath]
Until Layerpath appears in Claude's connector directory, add it using its URL. Once listed, you can connect from the directory using the same sign-in process.
**On Claude Free, Pro, and Max**
1. In Claude at claude.ai, open **Customize**, then **Connectors**.
2. Select **Add**, then **Add custom connector** at claude.ai.
3. Enter `Layerpath` as the name and `https://agent.layerpath.com/mcp` as the URL. Follow Claude's instructions to add the connector.
4. Select **Connect** next to Layerpath.
5. Sign in to Layerpath with Google or your email and password, or create an account.
6. Select the workspace where tours created in Claude will be saved, then select **Allow**. You'll return to Claude.
**On Claude Team and Enterprise**
A Claude organization Owner adds Layerpath through **Organization settings** → **Connectors** at claude.ai. Each person then opens **Customize** → **Connectors**, selects **Connect** next to Layerpath, and signs in to Layerpath.
**In a chat**
At claude.ai, select **+**, open **Connectors**, and turn on Layerpath.
**In Claude Code**
In Claude Code, type `/mcp`, choose Layerpath, then select **Authenticate** to sign in. If you've connected Layerpath at claude.ai, it appears as **claude.ai Layerpath**.
## Make a tour [#make-a-tour]
Ask Claude to create a tour. For example:
* Make a product tour from this Loom video: (your link)
* Make a tour of our signup flow
With a Loom link, Layerpath starts creating your tour directly. Otherwise, it shows a **Create with Agent** box in the chat. Drop screenshots or a screen recording on **Drop files here**, or select it to choose files. The tour starts when the files are added. If it doesn't, select the arrow to start it.
Files attached to your Claude message don't reach Layerpath. Add the same files in the **Create with Agent** box instead. Claude tells you when this happens.
Layerpath shows creation progress, then the tour's steps. Creating a tour usually takes a few minutes. Longer videos may take longer.
## What you can ask [#what-you-can-ask]
* Create a tour from screenshots, a screen recording, or a Loom link.
* Shorten steps, clarify instructions, or change their tone.
* Translate a tour in place or create a translated copy.
* Blur private information, such as names or email addresses.
* Move, merge, or remove steps, or change the intro or outro.
* Undo a change.
* Find an earlier tour by name and open it.
* Get a tour's share link or embed link.
* Export a tour as a video.
Select **Edit** on a tour to open it in the editor. There, you can select and edit a step or choose **Play** to preview the tour.
Claude may ask for permission before Layerpath makes a change. For some edits, Layerpath shows a proposed plan first. Select **Go ahead** to apply it, or reply in the chat to revise the plan.
**What Layerpath doesn't do in Claude**
* Layerpath can't generate AI voiceovers in Claude. Add voiceover to your tour in Layerpath at app.layerpath.com.
* Tours with voiceover can't be translated in Claude.
* Layerpath can't create tours from YouTube links in Claude. Use a Loom link, or add screenshots or a screen recording in the **Create with Agent** box.
* Editing a step's text removes its existing voiceover so the audio doesn't contradict the text. Video exports retain voiceover on unchanged steps.
* Layerpath needs screens from existing software. It can't design screens for software that hasn't been built.
## Where your tours are saved [#where-your-tours-are-saved]
Tours created in Claude are saved in the Layerpath workspace you selected during connection. They appear under your account alongside your other tours at app.layerpath.com, where workspace members can open them.
Layerpath publishes new tours at a share link when they're ready. Anyone with the link can open them. For an unpublished tour, asking Claude for a share link or selecting **Copy link** publishes it.
To switch workspaces, disconnect Layerpath in Claude, reconnect it, and select the other workspace.
## Who can change a tour [#who-can-change-a-tour]
You can edit tours you created. Owners and Admins can edit any tour in the workspace. **Undo** reverses changes made through Claude, not edits made directly in Layerpath.
## Disconnect Layerpath [#disconnect-layerpath]
**In Claude:** open **Customize**, then **Connectors**, and remove Layerpath. On Claude Team and Enterprise, an Owner can remove Layerpath for everyone through **Organization settings**, then **Connectors**, at claude.ai.
**In Layerpath:** Owners and Admins can open **Settings**, then **Integrations**. The Claude card lists connected members, each with a **Disconnect** option. Once disconnected, that person must reconnect to use Layerpath in Claude.
Tours already created remain in the workspace.
## If something goes wrong [#if-something-goes-wrong]
* **"Your Layerpath session has expired":** reconnect Layerpath in Claude.
* **"You're no longer a member of this Layerpath workspace":** reconnect Layerpath and select a workspace you belong to.
* **"This tour belongs to another of your Layerpath workspaces":** reconnect Layerpath and select the workspace containing the tour.
* **"Only the tour's creator or an Owner or Admin in (your workspace) can edit it":** ask the tour's creator or a workspace Owner or Admin to make the change.
* **You can't select a workspace when connecting:** if you're a Viewer, ask an Owner or Admin to change your role to Editor.
* **A share link doesn't open:** the tour may be unpublished. Ask Claude for its share link to publish it.
## Limits [#limits]
* Screenshots must be PNG or JPEG files, up to 15 MB each.
* Screen recordings can be up to 200 MB and 10 minutes long.
* Loom videos can be up to 10 minutes long.
* Claude Code and Codex can't send screenshots or recordings to Layerpath. Use a Loom link, or create a tour at app.layerpath.com and ask for it by name to edit it.
* AI can make mistakes. Review your tour before sharing it.
## Privacy [#privacy]
Layerpath receives each request as Claude words it, including your link, the relevant tour and step, and text to save. Claude may include details from your chat or its memory. Layerpath can't read your other Claude conversations, your Claude memory, or files attached to the chat. Screens added to the **Create with Agent** box go directly to Layerpath. Layerpath doesn't use any of this to train AI models.
Read how Layerpath handles your data: [https://www.layerpath.com/privacy](https://www.layerpath.com/privacy)
## Get help [#get-help]
Email [support@layerpath.com](mailto:support@layerpath.com). Include that you were using Claude, the tour name or share link, your request, what happened, and the time with your time zone.
---
# Use Layerpath in Slack
Canonical URL: /integrations/slack
Description: Create interactive tours in Slack from screenshots, screen recordings, or Loom or YouTube links. Edit them in the thread.
Layerpath turns real software screens into interactive tours with hotspots and callouts. In Slack, mention @Layerpath with screenshots, a screen recording, or a Loom or YouTube link. Layerpath creates the tour in the thread. Reply there to edit it or request a video.
## Add Layerpath to Slack [#add-layerpath-to-slack]
An Owner or Admin of a Layerpath workspace can connect it to Slack for everyone in that Slack workspace.
1. Open [https://app.layerpath.com/slack/install](https://app.layerpath.com/slack/install) and sign in to Layerpath or create an account.
2. Select the Layerpath workspace where tours created in Slack will be saved. You must be an Owner or Admin of that workspace.
3. Select **Continue to Slack**, then authorize Layerpath in Slack.
Layerpath sends you a direct message in Slack when connected. Your Slack workspace may require an admin to approve the app first.
Owners and Admins can also connect from Layerpath. Open **Settings**, then **Integrations**, and select **Add to Slack**.
Add Layerpath to a channel with `/invite @Layerpath`.
* Each Slack workspace connects to one Layerpath workspace.
* With Slack Enterprise Grid, add Layerpath to each Slack workspace separately.
## Make a tour [#make-a-tour]
Mention @Layerpath in a channel or thread and attach screenshots, a screen recording, or a Loom or YouTube link. You can also message Layerpath directly.
Layerpath shows progress in the thread, then posts a card with a thumbnail of the tour's intro. The thumbnail doesn't play in the thread. Select the card to open Slack's side panel. Select **Open tour** to open the tour or **Make video** to export it as a video. Reply in the same thread to make changes. The link stays the same.
Layerpath can also use screenshots and videos shared earlier in the same thread. Creating a tour usually takes a few minutes. Longer videos may take longer.
## What you can ask [#what-you-can-ask]
* Create a tour from screenshots, a screen recording, or a Loom or YouTube link.
* Shorten steps, add detail, or change their tone.
* Translate a tour or individual step.
* Add, regenerate, or remove voiceover, or change the voice.
* Blur private information, such as names or email addresses.
* Move, merge, hide, or remove steps. Change the intro, outro, tour name, or hotspot appearance.
* Undo the last change.
* Export a tour as a video. Layerpath posts the video in the thread when it's ready.
* Show your latest tours.
For some changes, Layerpath shows you the proposed changes first. It waits for your confirmation in the thread before applying them.
Layerpath needs screens from existing software. It doesn't design screens for software that hasn't been built.
## Who can use it [#who-can-use-it]
Members of your Slack workspace can use Layerpath. Guests and people from other companies in shared channels can't.
Layerpath responds to mentions, direct messages, and replies in threads it's working in. It doesn't respond when a reply mentions someone else.
## Where your tours are saved [#where-your-tours-are-saved]
Tours created in Slack are saved in the connected Layerpath workspace. They appear under **Layerpath Agent**, a member Layerpath adds to that workspace.
Layerpath publishes each new tour at a share link when it's ready. Anyone with the link can open it.
## Who can change a tour [#who-can-change-a-tour]
* In Slack, any member of your Slack workspace can edit a tour created in Slack, including tours someone else requested.
* In Layerpath, workspace Owners, Admins, and Editors can edit tours created in Slack. In Slack's side panel for the tour, open the **⋯** menu and select **Edit in Layerpath** to open the tour in the editor.
* Tours created directly in Layerpath can't be edited from Slack.
## Disconnect Layerpath [#disconnect-layerpath]
**In Layerpath:** Owners and Admins can open **Settings** → **Integrations** and select **Disconnect** on the Slack card. After confirmation, Layerpath removes the app from Slack.
**In Slack:** remove the Layerpath app from your Slack workspace settings.
1. In Slack, select your workspace name in the sidebar.
2. Open **Tools & settings**, then **Manage apps**.
3. Open **Installed apps**, find Layerpath, and select **App details**.
4. Open **Configuration** and select **Remove app**.
After disconnection, members of your Slack workspace can no longer create tours with Layerpath. Existing tours remain in your Layerpath workspace.
## If something goes wrong [#if-something-goes-wrong]
* **Layerpath didn't respond:** mention @Layerpath or reply in a thread Layerpath is working in. For channels, invite Layerpath with `/invite @Layerpath`.
* **"This tour was created in Layerpath, so I can't edit it from Slack":** open the tour in Layerpath with **Edit in Layerpath**.
* **"Layerpath isn't available to guests":** ask a member of your Slack workspace to use Layerpath.
* **"Layerpath is no longer connected to (your workspace)":** ask an Owner or Admin to reconnect Layerpath at [https://app.layerpath.com/slack/install](https://app.layerpath.com/slack/install).
* **"Only an Owner or Admin of a Layerpath workspace can connect it to Slack":** ask an Owner or Admin of your Layerpath workspace to connect it to Slack.
* **Your Slack workspace is already connected to another Layerpath workspace:** contact the person who connected it, or disconnect Layerpath from Slack before connecting another workspace.
## Limits [#limits]
* Layerpath accepts up to 40 screenshots and videos from a thread, with a combined limit of 250 MB. Each file can be up to 200 MB.
* Loom and YouTube videos can be up to 10 minutes long.
* AI can make mistakes. Review your tour before sharing it.
## Privacy [#privacy]
When you mention Layerpath in a thread, it reads the thread's messages and files. It also reads subsequent replies, including conversations between other members, to understand the thread's context. Layerpath uses those messages to write tour content only when you ask it to.
Read how Layerpath handles your data: [https://www.layerpath.com/privacy](https://www.layerpath.com/privacy)
## Get help [#get-help]
Email [support@layerpath.com](mailto:support@layerpath.com). Include your Slack workspace name, a thread link (open the message menu and select **Copy link**), your request, what happened, and the time with your time zone.
---
# Hotspots, areas, and callouts
Canonical URL: /interactivity/hotspots-areas-callouts
Description: Guide attention with a point, region, or message, and add an authored zoom when a step needs focused framing.
Point to the right target, explain the next action, and keep the guidance readable at the size viewers will use.
## Choose the right guidance [#choose-the-right-guidance]
| Type | Use it when |
| ----------- | --------------------------------------------------------- |
| **Hotspot** | One precise point is the target. |
| **Area** | A larger region is the target. |
| **Callout** | The viewer needs a guidance message attached to a target. |
Use the smallest target that remains clear. The guidance should name the action or expected result without repeating everything already visible on screen.
> **Keep guidance focused.** Good guidance answers one question: what should the viewer do next? Use a hotspot or area when the screen is already clear. Use a callout when the viewer needs more context.
## Add and refine guidance [#add-and-refine-guidance]
Choose the step, open **Insert**, then choose **Hotspot**. Place it on the control or region the viewer should use.
Select the guidance, open its type menu, then choose **Hotspot**, **Area**, or **Callout**. Confirm that the guidance still points to the intended target.
Use a direct action such as “Choose Add member.” Include the result only when it helps the viewer confirm the action.
Use the contextual toolbar to format text, add a link when appropriate, choose readable colors, and open the position menu.
Choose **Auto** first. If the message covers the target or important content, choose **Top**, **Left**, **Right**, or **Bottom**, then check the full step.
Choose **Preview**. Check that the target is accurate, the message does not cover important content, and the next action is obvious on desktop and mobile.
## Add an authored zoom [#add-an-authored-zoom]
Use an authored zoom when one step needs a specific focus region. Add the zoom to that step, select it, then choose **Preview Zoom** to check the framing.
Each step supports one authored zoom. Delete it when the framing no longer helps. Authored zoom is separate from the tour-level **Auto zoom** setting.
* Guidance types are not step types.
* A guidance position that works on desktop can cover content on a smaller screen.
* This page does not cover masks, embeds, text overlays, arrows, buttons, or images.
* Preview the complete tour after changing guidance or Zoom.
## When guidance is not what you expect [#when-guidance-is-not-what-you-expect]
### The message covers the target [#the-message-covers-the-target]
Choose **Auto** or move the message to another side. Shorten the instruction if it still takes too much space.
### The target moves away from the control [#the-target-moves-away-from-the-control]
Return to the editor, select the guidance, and place it again. Preview at desktop and mobile sizes.
### Zoom makes the step harder to understand [#zoom-makes-the-step-harder-to-understand]
Preview the Zoom, then delete it if the wider context is more useful than the focus region.
## What to do next [#what-to-do-next]
Configure tour playback
---
# What's new
Canonical URL: /whats-new
Description: Recent Layerpath changes and the Help Docs pages that explain them.
Find recent changes that affect how you use Layerpath or follow these Help Docs. Product changes and documentation changes are labeled separately.
Start with images, video, library media, or a Loom or YouTube link. Agent prepares an initial tour for your review.
Keep the tour, selected step, and editing controls together on the canvas.
Set up identity and voice guidance before you create a new project.
## August 26, 2026 [#august-26-2026]
### We rebuilt the editor [#we-rebuilt-the-editor]
The rebuilt editor keeps the tour, the selected step, and the controls you need close together on one continuous, scrolling canvas.
* **Edit in place.** Rewrite callouts where they appear, position hotspots against their targets, and adjust layouts without leaving the selected step.
* **See the whole story.** Move through the tour, then add, reorder, duplicate, hide, or delete steps from one place.
* **Refine every step.** Add voiceover and adjust layouts, backgrounds, zoom, and masking without losing context.
* **Shape it together.** See teammates on the canvas and follow changes as they happen.
[Read the editor release post](https://www.layerpath.com/resources/blog/editor).
## August 12, 2026 [#august-12-2026]
### Brand memory Beta [#brand-memory-beta]
**Brand memory** is a Beta workspace feature for identity and voice guidance. Set it up from a website, a brand PDF, or your own assets. When it is configured, it can guide the identity and wording Layerpath prepares for new projects.
Review the generated identity, wording, visual choices, and voice settings before you share a project. Existing projects keep the Brand memory captured when they were created.
[Set up Brand memory](/brand-memory/set-up-brand-memory) explains the setup paths. [How Brand memory affects creation](/brand-memory/how-it-affects-creation) explains what it can guide.
## July 16, 2026 [#july-16-2026]
### Create with Agent [#create-with-agent]
**Create with Agent** is the creation path for material you already have. Add images or video, reuse saved library media, or use a Loom or YouTube link, then describe what viewers should accomplish. Agent prepares an initial tour that you review and refine in the editor.
The complete procedure, source choices, and Advanced options are in [Create a project with Agent](/create/create-with-agent).
* **Record your screen** remains the separate path for capturing a browser workflow.
* Use the source choices and Advanced options shown in your current workspace.
## July 15, 2026 [#july-15-2026]
### New Help Docs for core editing tasks [#new-help-docs-for-core-editing-tasks]
The Help Docs now include focused task pages for:
* [Managing steps](/edit/manage-steps).
* [Layouts, backgrounds, and core visual settings](/edit/layouts-and-backgrounds).
* [Navigation, zoom, motion, autoplay, and cursor behavior](/edit/tour-playback-settings).
* [Hotspots, areas, and callouts](/interactivity/hotspots-areas-callouts).
These pages explain controls that are already available in the editor.