#0001
`Airweave` launches unified app search for agents via API and `MCP`
60radar
AirweaveSearch layer — unified indexing across SaaS and DBs
Instead of wiring each SaaS tool separately, you get one retrieval layer over apps and databases. Useful when agents fail on missing context more than tool execution.
- Content is crawled via source APIs, then normalized, chunked, entity-linked, and indexed in a vector store plus
Postgresmetadata. - Retrieval runs semantic search and
BM25in parallel, fuses withRRF, adds recency bias, then re-ranks for chunk-level citations or synthesized answers. - The same surface ships as
REST, Python/TS SDKs, andMCP, so one integration can serve app code and agent tooling together. Temporalhandles sync orchestration like pagination, rate limits, schedules, and change detection, which cuts a lot of glue code for live data ingestion.
Source: github.com/airweave-ai/airweaveRead original →