I set this up a few months ago based on asciimoo's comments on HN, and barely used it at first, but I realized not too long ago that it could be a pretty useful research tool for one of my hobbies (award travel), that revolves around being in the know around various concepts and quirks.
I scraped and imported posts from the blogs I regularly reference for award travel, then hooked it up to OpenCode/Codex as an MCP server and used that corpus for research on those topics. So I can ask things like "has anyone ever mentioned running into this problem before?" [1]
If you have a hobby or working situation that requires you to regularly reference a core set of websites or reference materials, Hister provides almost all the tools out of the box to start a search engine against it. The default datasets they promote include the Python Stlib, MDN and RFC corpus, as an example. [2]
Thank you so much for building this! It's a really awesome piece of work and I'm grateful for your work. I'd love to sponsor you on Github in the near future.
The only thing that I think would be interesting to see is native support for crawling via a sitemap.xml instead of recursively. I worked around this by implementing a basic scraper that fetched pages exclusively from the sitemap.xml to add into Hister.
I think you're already aware of this, but I also experienced some data loss during the import because I was running a concurrent reindex. I clocked it pretty quickly so I didn't think too much of it. [1]
reply