Hister
A personal search engine over your pages and files with MCP access: the agent searches your personal index and the data stays with you
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Indexes your pages and files on a local server
- Semantic search sends text to a chosen embeddings endpoint
Install
Manual install
brew install histerDocker and Nix are also available. After install, start the server and open http://127.0.0.1:4433.
This is third-party code. Review the repository files before installing.
What it does
Hister is a private search engine over the pages you visit and the files you keep. It indexes their full contents on your own server, and you can search from the web interface, a terminal, the command line or an AI assistant over MCP. Pages enter the index through a browser extension for Firefox and Chrome, plus history and bookmark import and a site crawler. By default there is no telemetry and no cloud sync; content goes only to the Hister server you configure. Optional semantic search sends text to an embeddings endpoint you choose.
Who it is for. For people who keep a personal library of what they have read and want an agent to search it without handing data to an external search provider.
Good fit when
- You want the agent to find information across your saved pages and files
- You want private search without a third-party provider
- You want one index for the web interface, terminal and agent
Not a fit when
- You cannot run a local server and index
- You need open-web search rather than your personal index
Example request
Find the page in my index where I read about cache setup and quote the fragmentLimitations
Hister must be self-hosted: the server has to run while indexing and searching. The index is filled through the browser extension, import or crawler, so an empty server finds nothing. Details of connecting an MCP client to your server are in the documentation. Semantic search sends text to a chosen embeddings endpoint, which matters for private data.
How to disable. Remove the Hister server entry from the MCP client config, stop the Hister server and, if you wish, delete its data and the browser extension.
MCP
- Transport
- http
- Authentication
- not required
Security check
- Indexes your pages and files on a local server
- Semantic search sends text to a chosen embeddings endpoint
README in short
The README describes Hister as a private search engine over your pages and files that indexes full contents. You can search from the browser, terminal, command line and an AI assistant over MCP. Install via Homebrew, Docker or Nix; the server runs locally, filled through a browser extension, history and bookmark import or a crawler. It notes multi-user mode, no telemetry or cloud sync by default and optional semantic search through a chosen embeddings endpoint. The backend is Go, AGPL-3.0 licensed.
FAQ
Where is the data stored?
On your Hister server: no telemetry and no cloud sync by default, and the extension sends content only to the server you configure.
What is needed for search to work?
A running server and a filled index: add pages via the extension, history and bookmark import or the crawler.
Related
CLI, skill and Python library for browser control: the agent clicks, fills forms and reads pages over CDP
A skill that gathers the last 30 days of discussion on a topic from Reddit, X, YouTube, HN, Polymarket and GitHub into one brief
The Chrome DevTools team's official MCP server: the agent drives a live Chrome, reads network and console, and records performance traces
A fast Rust CLI for agent browser automation: accessibility snapshots with element refs, an MCP server and skills