What it does
This MCP server integrates Airweave, a unified context retrieval platform, with Claude. Airweave sits between your data sources—apps, databases, documents—and AI systems, syncing and indexing data from 50+ integrations. The MCP server allows Claude to search Airweave collections and retrieve grounded context on demand, whether from a cloud instance or self-hosted deployment.
Who it's for
AI engineers and full-stack developers who have deployed Airweave (cloud or Docker) and want Claude to query its unified retrieval layer during conversations. Useful for teams building agentic systems that need current context from connected apps like Airtable, Asana, or databases without rebuilding data pipelines for each integration.
Common use cases
- Retrieve context from Airweave collections during Claude conversations to ground responses in current data
- Query multiple data sources (CRMs, project management, databases) through a single unified search interface
- Build multi-turn agent workflows where Claude fetches fresh context from Airweave on each turn
- Supplement Claude's knowledge with real-time data from connected business apps
- Prototype retrieval-augmented generation without managing fragile data pipelines
Setup pitfalls
- Requires a running Airweave instance (cloud at
app.airweave.aior self-hosted via Docker) - Self-hosted setup needs Docker, docker-compose, and 9 ports available (8080, 8001, 5432, 6333, 6379, 7233, 8081, 8088, plus ephemeral ports)
- Airweave API credentials or auth tokens must be configured; verify connectivity before starting the MCP server
- Reads and writes to filesystem; ensure sandbox permissions allow this
- Network calls to Airweave backend and integrated apps—firewall or proxy rules must permit outbound traffic