tribal-memory/tribal

Overview

Self-hosted semantic memory system built in Rust that uses Postgres with pgvector for vector storage. It captures and retrieves an engineering team's tacit knowledge through the Model Context Protocol (MCP), allowing queries over internal know-how without sending data to external services.

Best for

Best for
Engineering teams that want to preserve tacit knowledge locally with semantic search over their own infrastructure.

Use cases

• Store and search internal design decisions and architectural rationale
• Retrieve past debugging solutions from team chat logs or code comments
• Query recurring patterns and workarounds from private codebases

How to use

Install

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/tribal-memory/tribal/releases/latest/download/tribal-installer.sh | sh

Tools exposed

• tribal

Example client config

{"servers": [{"kind": "tribal-memory/tribal", "url": "http://localhost:8080"}]}

Notes

Self-hosted semantic memory system built in Rust that uses Postgres with pgvector for vector storage. It captures and retrieves an engineering team’s tacit knowledge through the Model Context Protocol (MCP), allowing queries over internal know-how without sending data to external services.

5 stars on GitHub. Last updated 2026-06-14.

Use cases

• Store and search internal design decisions and architectural rationale
• Retrieve past debugging solutions from team chat logs or code comments
• Query recurring patterns and workarounds from private codebases

Pros

• Fully self-hosted, keeping sensitive knowledge on your own infrastructure
• Built on common stack (Postgres + pgvector) for easy integration
• Written in Rust for performance and safety

Cons

• Requires setting up and maintaining an MCP server infrastructure
• Very early stage with only 5 GitHub stars and limited community
• Needs separate AI model access to make full use of the semantic memory

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.