Quivr is a free, open-source engine that turns continuous streams of content — news feeds, mailboxes, posts and documents — into durable search and monitoring. Maintained by The Vibe Company, the project has earned more than 39,000 stars on GitHub and is under active development, with its latest updates in 2026. It is deliberately an engine, not a finished app: your product calls its API, and Quivr handles ingestion, indexing and alerting underneath.
Sources such as RSS feeds, Microsoft 365 mailboxes and X lists are collected on a schedule, and every document is stored with its full version history. Background plugins prepare text for both keyword and meaning-based search, rank results and check new arrivals against saved alerts; when something matches, a signed webhook notifies your application with the reason. AI agents can search Quivr through MCP and quote passages with their sources. Processing is durable by design — an outage delays work but never loses a confirmed document. Idempotency keys and tracked receipts make submissions safe to retry, which matters when feeds and partner systems push documents around the clock.
Quivr is free and open source under the MIT license. You run the engine on your own servers next to your application; there is no hosted tier or per-document fee from the project itself.
Quivr is for developers building newsroom tools, media-monitoring services, research assistants or any product where content keeps arriving and users need to search it, follow topics and get alerted — without building the ingestion pipeline from scratch.