Hister: A self-hosted, full-content search index for pages and files you keep

Hister is an open-source (AGPLv3) search engine you run yourself, designed to index the full text of web pages you visit and local files you keep rather than just their titles, URLs, or bookmarks. Content is collected through a browser extension, local folder watchers, browser-history imports, or a site crawler, then extracted and indexed on a server you control. The pitch is durable personal recall: find something useful once, and it stays searchable in an index that lives on your own machine.

The privacy model is the main selling point. The server holds the index, stored page content, and indexing rules locally, with no telemetry and no required cloud service—it doesn’t report what you search. A full setup can run on a single local machine as one binary, while shared deployments support multiple users with SQLite or PostgreSQL and ship via Docker or Nix. Optional features that involve external connections, such as semantic search via a configurable embeddings endpoint or favicon fetching, are left to the user to enable and point wherever they choose.

Search is exposed through a web UI, terminal client, CLI, HTTP API, and an MCP server, the last of which lets AI assistants retrieve indexed content directly. The query language supports field filters, quoted phrases, wildcards, negation, date ranges, and user-defined aliases, with skip and priority rules controlling what gets indexed and how it ranks, plus content versioning to retain earlier document states. For technical users wary of cloud-based note and bookmarking tools, Hister offers a self-owned alternative that keeps the actual content—not just links—searchable and under local control.