Menu

Turn video walkthroughs into clear documentation, clear specs, product feedback, and help center guides.

Record a workflow. Turn it into a help guide or spec you can edit and share. Open source, with your choice of AI.

InstallView on GitHub

Apache-2.0 license Local or Cloud AI v0.3.0 prerelease

1 / 2Your recording, activity, and document — in one workspace.View full size
Open sourceDesktop captureMarkdown exportLocally stored media

Capture your workflow

Record a screen, window, or region, with optional microphone and click capture.

Keep the useful details

Review the video, click screenshots, and transcript together. Edit what needs a correction.

Make it documentation

Generate a help guide or spec, refine the Markdown, and take it wherever you work.

Choose your AI

Use a local Ollama model or your preferred cloud provider. Your media is stored locally.

A simple path from screen to document

Do the work once.
Make it easy to follow.

  1. 01

    Show what you mean.

    Pick a source and record a walkthrough. Explain the process, demonstrate a feature, or talk through an improvement.

  2. 02

    Review what happened.

    Replay the video, seek from activity timestamps, and correct the transcript before turning it into a document.

  3. 03

    Choose a flow. Make it yours.

    Select Help Guide, Spec Document, or your own instructions. Generate, edit, and export Markdown.

Help guides

Onboarding, product instructions, and the everyday workflows worth writing down.

Improvement specs

Turn a product walkthrough and your observations into a focused document to refine.

Your tools. Your choice.

Keep the capture local.
Choose how you use AI.

Recording and video processing work without a cloud AI provider. When you're ready to generate, choose the model that fits your workflow.

Read about AI and privacy
LOCAL

Ollama

Use a vision-capable model on your computer. A remote endpoint receives the supplied context.

CLOUD

OpenAI, Anthropic, Google

Bring a provider key. The context needed for click analysis or document generation goes to your selected provider.

A few things to know.

Is Path free and open source?

Path is open source under the Apache-2.0 license. You can explore, run, and contribute to the source on GitHub. If you choose a cloud AI provider, its usage costs and terms apply.

How do I get started?

On Windows, download the installer using the Install button above. For source setup, follow the repository guide. Running from source requires Node.js 22.14 or newer and npm 10 or newer. Some native dependencies may need C++ build tools.

Read the setup guide
Which platforms can I use?

Windows is the primary development target. macOS packaging is configured, but native runtime and permissions still need validation. Linux does not currently have an installer target.

Does everything stay on my computer?

Recordings, screenshots, and transcripts are stored locally. Capture and video processing do not require cloud AI. If you select a cloud provider or remote Ollama endpoint, the context needed for that operation is sent there. Cloud document generation does not send raw video or audio.

What should I know before using it?

Generated documents need your review. Markdown drafts are not autosaved, so copy or export them before switching recordings or closing Path. System audio and user-facing pause and resume are not yet available.

See current limitations

Open source by Automated.

Use it. Inspect it. Improve it.

Explore the source, try Path, and help shape what comes next.

Explore the repository