Hister

Hister is a self-hosted personal search engine that indexes the full contents of the web pages you visit and the files you keep on your own machine. It is aimed at people who lose things they have already read: someone who remembers a paragraph but not the site it was on, a developer who wants a queryable archive reachable from the terminal, or anyone who would rather keep that index locally than hand their browsing history to a cloud service. It is written in Go by asciimoo, released under AGPL-3.0, and has gathered roughly 4,900 GitHub stars since the repository was created in January 2026 — including, as the video notes, about 800 of them in a single day.

What it does

Ordinary search engines index the public web. They are not much help when the thing you want is a page you personally opened three weeks ago, or a document sitting in a folder on your laptop. Hister fills that gap by indexing the full text of both, so the search box covers your own reading history rather than the internet at large.

It exposes that index three ways:

• a local web interface, the screenshot of which leads the README;
• the command line, for people who would rather not leave the terminal;
• an MCP server, so an AI assistant can query your indexed pages and files as a tool.

Content gets in through several routes. A browser extension captures pages as you visit them, and you can also import an existing browser history, index local directories, or import individual files after the fact — so the archive does not have to start empty on the day you install it. The project's own GitHub topics list semantic search alongside plain indexing, and there is a hosted demo at demo.hister.org if you want to try the interface before downloading anything.

How it works

Hister runs as a single local binary. You start it with hister listen, and it serves on http://127.0.0.1:4433. That server process has to stay running while you use Hister — it is what receives pages from the browser extension and answers searches. Nothing leaves the machine unless you arrange for it to; the README frames the default as a local personal setup requiring no configuration at all.

The browser extension is published for both Firefox and Chrome and is installed from the usual add-on stores. Indexing is opt-in per page in the sense that the extension has to be enabled where you want capture to happen, and the documentation covers how to narrow down what Hister takes in.

Getting started

The quickstart in the README is short enough to follow in a couple of minutes:

1. Download the binary for your platform from the latest release and rename it to hister (or hister.exe on Windows).
2. On Linux or macOS, make it executable with chmod +x hister.
3. Run ./hister listen, or .\hister.exe listen in PowerShell on Windows, and leave the terminal open.
4. Open http://127.0.0.1:4433 and install the Firefox or Chrome extension.
5. Visit a page with the extension on, then search Hister for a phrase from it.

If you would rather not manage a downloaded binary, the project also documents Homebrew (brew install hister), Docker, and Nix installs. Fuller documentation, including the import and configuration guides, lives at hister.org/docs.

When to use it / when not

Hister makes sense when your problem is re-finding, not discovering. If you read widely and regularly hit the wall of "I know I saw this somewhere", a full-text index of your own history is exactly the right shape of tool, and having it answer over MCP means an assistant can dig through that archive on your behalf instead of guessing from memory.

It is less useful if you want to search things you have never seen — Hister only knows what it has indexed, so a fresh install is empty until you import history or files. It also expects a process to be running; this is a service you host, however lightly, not a passive browser feature. And because it captures what you browse, the index itself becomes sensitive: keeping it on your own hardware is the point, but it is still a database of your reading that deserves the same care as any other. Finally, the AGPL-3.0 license is worth reading before you build a product on top of it, though it poses no problem for personal use.

Hister is worth taking seriously if you treat your own reading as a resource rather than a stream you let go by. It is a small, self-contained Go program that turns browsing history and local files into something you can actually query, from a browser tab, a shell, or an AI assistant, without a hosted account in the middle. The project is young — created at the start of 2026 and actively pushed through September — so expect it to keep moving, but the core idea is already the sort of thing that is hard to give up once your archive has a few months in it.