About Wayfinder
Wayfinder is a local-first desktop app for reviewing how AI-assisted project work evolves, aimed at people working across coding, research, writing, and design. It connects goals, branches, failures, decisions, validation, and artifacts in a source-linked project history.
Key features
- 01Local project history maps
- 02Separate voyages by goal
- 03Visible branches and failures
- 04Replay recorded file changes
- 05Source-linked session details
Use cases
- 01Reviewing why an approach changed
- 02Comparing attempts across sessions
- 03Tracing validation and file edits
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEWayfinder
See how your work with AI actually evolved.
Wayfinder is a local-first desktop app that turns work with AI into a visual history of how a project reached its result. Goals, alternative attempts, failures, decisions, validation, and artifacts stay connected instead of disappearing into separate chats.
Wayfinder is designed for AI collaboration across coding, research, writing, design, and other project work. The current early-access collectors support Codex and Claude Code.
Download Wayfinder · Website · Get started · AI collaboration history · Integrations · Updates · Security · Privacy · Report an issue
No account. No telemetry. No cloud sync.
Why Wayfinder
AI tools are good at producing the next answer. They are less useful at showing why one approach worked, why another failed, or which evidence changed the direction of a project.
Wayfinder keeps that path reviewable:
- One project map across AI tools.
- Separate voyages for unrelated goals in the same workspace.
- Visible branches and failures instead of a cleaned-up success story.
- Real file-change replay when the source session contains an edit.
- Source-linked details for conversations, tool activity, files, and validation.
How It Works
- Install and open the desktop app.
- Keep working normally with your AI tools.
- Return to Wayfinder to review how the project evolved.
Related sessions are grouped locally. Every recorded turn retains its original source, and Wayfinder does not invent missing history or historical diffs.
Current collector details:
Downloads
Wayfinder early access is available for:
| Platform | Build | Download |
|---|---|---|
| macOS | Apple Silicon | DMG |
| macOS | Intel | DMG |
| Windows | x64 | EXE |
Download the latest installers and checksums from GitHub Releases. The current early-access builds are not store-signed; read the install guide before first launch.
Local By Default
Wayfinder stores normalized history, collection cursors, and project maps under
~/.wayfinder. There is no Wayfinder account, analytics service, advertising,
cloud sync, or cloud analysis in the current build.
Read PRIVACY.md before using confidential projects.
Current Scope
- Standalone desktop application for macOS and Windows.
- AI collaboration model spanning conversations, research, writing, design, coding, and other project work.
- Current public early-access collection from Codex and Claude Code.
- Project-level voyage grouping, branches, source details, and file-change replay.
- Early access: data formats and grouping behavior may continue to evolve.
Historical plugins, MCP bundles, IDE extensions, package-manager taps, and standalone CLI artifacts are not current product surfaces.
Development
npm ci
npm run check
npm run companion:prepare
npm run companion:build
The test suite covers collection, voyage grouping, the desktop renderer, trackpad pan and zoom, detail inspection, and narrow-screen behavior.
Read CONTRIBUTING.md before opening a pull request. Public copy and screenshot rules live in docs/PUBLIC-PRESENTATION.md.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.