Your data, one interface.

A local-first MCP server that syncs work tools and connects live data sources into a single searchable, queryable index. Single binary, zero config, data never leaves your machine.

🔍

Full-Text Search

BM25 ranking with fuzzy matching. Search across all connected sources with relevance-scored results and highlighted snippets.

📊

SQL Analytics

Run analytical queries over columnar data. Filter, aggregate, and join across structured records from any source.

🌐

Graph Traversal

Navigate relationships between entities. Discover neighbors, find shortest paths, and explore connections across tables.

🔄

Work Tool Sync

Incremental sync from chat platforms, project tools, and knowledge bases. Only pulls changed data on subsequent runs.

Live Data Connectors

Query external databases and services in real time. Connectors translate SQL to native query languages on the fly.

📦

Single Binary, Zero Config

One binary, no external services. Data stays on your machine. Set TEIDELUM_DATA and run — that's it.

How it works

Data flows from your tools through a unified engine to AI agents via MCP.

Teidelum architecture diagram

Get started in seconds

# Clone and build
git clone https://github.com/TeideDB/teidelum.git
cd teidelum && cargo build --release

# Run the MCP server
./target/release/teidelum

# Add to your MCP client config
{ "mcpServers": { "teidelum": { "command": "./target/release/teidelum" } } }