Trek
TREK is a self-hosted travel and trip planning application built for collaborative itinerary management. It combines interactive maps, real-time multi-user synchronization, budget tracking, and packing list management in a single deployable unit. As of v3.1.2 (June 2026), it has 7.2k GitHub stars and 631 forks, making it one of the more actively maintained self-hosted travel tools available.
- Interactive maps via Leaflet and Mapbox GL with drag-and-drop day organization.
- Real-time WebSocket synchronization across collaborators.
- Budget tracking with expense splitting and reservation management.
- Packing lists and document attachment.
- PWA support for mobile installation without an app store.
- Single Sign-On via OIDC providers.
- 20+ language support.
- Built-in MCP server for AI assistant integration.
TREK ships with a built-in MCP server, which allows an AI assistant to query and update trip plans directly. This makes it a complementary tool to the Geospatial MCP for AI Planning technique: TREK provides the planning data store and map interface while an AI assistant handles distance calculations, stop sequencing, and place discovery via a maps API.
Node.js 22, NestJS 11, SQLite, React 19, Vite, TypeScript, Tailwind CSS, Leaflet, Docker. Deployed via Docker with JWT and OAuth 2.1 authentication.
AGPL v3. Self-hosting for personal or internal use is permitted. Modifications to the network service must be disclosed as open source.
Placed at Assess. TREK is well-maintained with significant community adoption and active releases. The built-in MCP server is a meaningful differentiator for AI-assisted travel planning workflows. No first-person production use yet. Trial requires deploying an instance and evaluating the MCP integration against a real itinerary planning session.
