Ohi, author here! Thanks for posting Hister. Feel free to A.M.A.
My first free software search project was Searx, a privacy respecting metasearch engine, but because of the limitations of the metasearch concept, I've decided to take a different approach.
Hister builds a personal search index from pages you visit, bookmarks, browser history, local files, and crawled websites. It stores extracted content with offline result previews, so information remains searchable even when the original page changes or disappears. It supports full text and semantic search, can run entirely on your own machine, and includes a web interface, command line tools, and an MCP endpoint for assistant integrations.
>Hister builds a personal search index from pages you visit, bookmarks, browser history, local files, and crawled websites.
Immediately interested and will check it out, thank you! I've wanted a "search stuff you've seen online" tool for a long time, but everything seems to be research-oriented or "archive but don't search" or some weird combination that means it's nigh useless to me. I've got decades of bookmarks and archives and I've kinda been stuck grepping them at best (it's rare but I do sometimes want a page I saw once three years ago and I love having that option), while hoping someone would build something better.
One question if ya don't mind, while I explore: any chance of singlefile support? Content-extraction is useful in lots of situations (e.g. wallabag) and it's a great default, but sometimes it fails and sometimes you really do want the page, relatively close to how it actually was. Singlefile does that much better than most, and it does so well enough (and manually-handle-able enough if needed) that I don't feel any desire to switch to WARCs or similar.
Though specifically I'm probably looking for something like "content-extract everything" + "key combo to save singlefile version too" + "upload singlefile archives to backfill / recover". Like 99% of the time content extraction is preferred, and I'm glad to see it... it's just not always enough, and having to go elsewhere for exceptions breaks a lot of the utility.
Exactly! I had the very same issues before Hister.
> One question if ya don't mind, while I explore: any chance of singlefile support?
Yes, partially. Hister can already import HTML files created by SingleFile, but there is no direct integration yet. In the longer term, I would like the SingleFile extension to be able to send snapshots directly to Hister.
I assume it content-extracts that on upload? I'd really like to move storage into hister too, if possible. That way you could also switch from an extracted view to a "full" view in the UI. Though I assume that'd be fairly simple to build later.
Overall I really like what I'm seeing, it ticks a lot of important boxes for me and it's pleasantly straightforward. Hopefully I'll find time to contribute!
> That way you could also switch from an extracted view to a "full" view in the UI.
It isn't even needed, we just need a SingleFile specific extractor (an interface in Hister to parse specific page content and provide custom previews) that provides the full original HTML for the preview panel.
I came here to mention https://github.com/gildas-lormeau/singlefile, I couldn't get it to do what I want so I built my own, for watched domains it pushes a copy of the serialized DOM to a local search database. For structured data, either extract and enrich in the browser or enrich on the server side.
Would Hister support this basic workflow? I'd love to retire my own software.
The next phase was going to move to a recording proxy.
Singlefile based web pack+download is exactly the next feature I need besides indexing URLs / browser tabs with[0], have it on my to-research for a looong time. I do have a server-side script that can be triggered on doc ingestion as a hook "if data/schema/tab is indexed and path is /ml/* then run script "fetch-url.sh" and file the result into {{doc.path}}" but being able to do that in-browser is more user-friendly.
Thanks you for making Hister. I have been using it since May and find it immensely useful. It's a piece of mind to know that I always will be able to easily find what I looked at. So I don't feel the need to bookmark things, which I used to do, but then never looked at anyways ;).
I'm hacking on a in-process DB on top of LMDB+Lance(for now, hilbert space kung-fu with a custom matryoshka embedding setup with separate spatial + temporal + internal and content derived anchors will replace that ~last-century~ last-year tech) + roaring bitmaps as the primary indexing engine with the same or similar goal[0] and will definitely deep-dive into yours.
I'm also trying to index users unstructured documents and workflows(tabs, emails, files, notes, identities etc)
- Organize them into semantically meaningful user or agent created context or directory-like virtual trees (the same photo of a nice kitchen may be surfaced under `/travel/barcelona` and `/arch/interieour/kitches`)
- ..where tree nodes are mapped to bitmaps - `/travel/barcelona` does a fast and cheap `travel` AND `barcelona`, want to "zoom-out" you just go one directory up to `/travel` and see all documents tagged with travel)
- You can use multiple timelines - extract that fancy md-converted en-wiki hf dataset into a wikipedia db dataset + timeline, tag your personal timeline as "personal" - wanna know the zeitgeist of your grandmothers birth date - search for it with timelines personal + wikipedia in layered mode and you'll get everything that happened or was happening during that time.
- You can have long-running stateful query sessions and refine your searches dynamically - search for "winter" and get all documents with a winter scenery or mentioning winter - refine with "nice view" then "laptop" - citing a recent example[1]
- Documents have relations that would be cumbersome to map in a virtual tree structure(worth an experiment due to the zoom-in/out you get with context bitmap trees though) - hence on top of the initial structure you can use graph edges(also powered by bitmaps - as most indexes are)
- All vector queries always run on top of a candidate set you get by the bitmap/bitmap-based filter algebra hence searching through 100k+ docs is usually pretty fast
This sounds very interesting. Wanted something similar, so going to check it out:)
One thing at first look, like a little feature request ;).
Personally I think it would be very nice if it was possible to index different sites to different search indexs. So you could separate different stuff, like job, different projects you have, and other stuff and then last everything else etc to there on search index db.
So it would be possible to have like "profiles" you could easy choose from. So in the addon, when you click on the hister icon on the toolbar there would be a list of profiles that you could choose where to index the site to. But also a setting to choose which domains and sites that always should index to one search index profils. And all sites that dont is added to a profile is added to the default index instead.
For me this would make it much more easy to find stuff and sort out what Im looking for (for me things get very "thing"/project based what I want to find), so having this..
1 - The "profiles" would become like an important filter. When you search you could easy mark one or more "profiles" which you would search from, and you remove alot of unwanted data automatically, specially when you index every site you visit by automatically.
2 - It would also make it easy if the db gets big over time to remove data that is less important later on and that you dont need to be index anymore.
3 - It would also make it easier to backup only the most important search index DB:s if the DB for the differnt search index:s where in different folders/files. I guess the "default index" that index all sites could easy get big, but there it would be alot of not important data. So would be nice to be able to easy just use a backup program and only backup the only important search databases, to save space.
I was thinking about making a "personal data search engine". I have almost 30GB of email archives and a bunch of Google Drive and offline files. I am thinking about downloading it all from the cloud and storing it on a large RAID array, with an indexing and search interface (and possibly a local MCP server).
Would Hister be suitable for this? Can it index mbox files? Would it handle this amount of data? Does it have a search API so I can build an MCP server?
The indexer it uses (Bleve) can handle millions of records according to their docs, but sure that Hister would be the best choice for this task. I'd probably use Meilisearch (https://www.meilisearch.com/).
> Can it index mbox files?
Not yet.
> Does it have a search API so I can build an MCP server?
Safari support is currently stalled in this PR: https://github.com/asciimoo/hister/pull/46. I do not have access to macOS, so help with development and testing would be very welcome.
You should integrate with Karakeep.app, it's only natural that you'd want both a search engine, and a nice "archive" and "article pretty view" features :)
What do you mean by integration? Hister already supports importing data from Karakeep, Linkding, Linkwarden, Readeck, Shaarli, and wallabag: https://hister.org/docs/import
It depends on what you consider intense browsing. An indexed document uses about 100KB on average because Hister stores the full original HTML for offline previews. If storage is a concern, you can disable offline previews, reducing the average document size significantly.
Not in the near term. Right now I am focused on developing Hister rather than operating a hosted service. There are already plenty of centralized hosted search engines, so my longer term interest is in federation and distributed search. I want to make the core system mature first.
> would've been great with a more liberal license
It depends on how do you define liberal. =] I chose AGPLv3+ because I want Hister to remain free software and available to their users.
I agree with AGPLv3, I don't really understand why people have an objection. All it prevents is someone taking your work and making a business out of it.
I used to be interested in the prophecies of Nostradamus a lifetime ago. You may be surprised to learn that “Hister” is a name mentioned in those prophecies that is commonly claimed by believers to refer to the future Adolf Hitler.
Academics usually take it to mean the Danube river instead, but in conspiracy/New Age contexts, the association with Hitler is prevalent, and mentioned in several pop culture works, including at least one feature film.
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]
I just finished setting this up yesterday and I'm kind of obsessed with it. I was previously a heavy user of Karakeep, but I hated forgetting to save something and losing it. I also think Hister's semantic search is a better solution than AI generated summaries and tags. Support for local docs is super cool too, I have it set up to index my org notes directory.
This is so cool! For the time being I’m still locked into notion for my handwritten knowledge base, but I love this for incorporating external information.
For the semantic search is there any chunking/processing that happens with the content or do you need to be diligent about having a large embedding context (and/or small content)?
I tried this out this week and liked it but really wish this project had some form of auth. Opening the contents of every page you’ve ever visited, even to the local network, is not the best idea.
Hister supports token based, password based, and OIDC/OAuth authentications with optional multi-user handling. Details about user handling can be found here: https://hister.org/docs/user-handling
It also has a "public mode" where anyone can search the indexed content, but only authenticated users can add or modify it.
It seemed like the public mode was the default when I set it up. If so, that’s a fairly dangerous default as keeping a “clean” history with no secrets leaked seems neigh impossible.
The default configuration binds only to localhost, and a fresh installation starts with an empty database/index. Could you clarify which specific attack surface you are concerned about in that scenario?
I’m not concerned about an attack scenario. I’m just saying that using the docker image, if someone (or their agent) isn’t careful, they could expose their browsing history publicly fairly easily. It might just be nice to default to at least a user and pass login rather than just wide open.
I think that the tells, if they are in fact legitimate, are negligible at best or subtle at worst. It looks like the site uses Tailwind, so if AI was involved it wasn't generating the design from scratch. If it did then I'm impressed because usually it sucks. This is alright.
Really? The website, you mean? :( I thought to myself, "what a cute, well designed webpage!" when I saw it. Maybe with time this site will stand out like those Claudeslop ones.
I was not using Wikipedia to get there. This is something I immediately recognized from research about World War II and possible prophetic fulfillment. I definitely wasn't playing a game when I came across this information either, but I get your drift.
Where it says that “Hister” is a variant of its Latin name.
I think the name is lovely with no reason to change. Surely there are few people who would connect this with silly medieval superstition. Perhaps you can take part in wiping out that association.
I've been on site site for years. I created an account just to comment on this thread. Holy crap, the only time I've ever heard 'Hister' was in a book I read on Nostradamus like 30 years ago.
Doh, sorry, the answer was coming from me who did not read the question properly. It currently cannot import browser bookmarks, but it is a good idea. Added to my TODO.
I love the idea overall, but something about browser extensions give me a bad taste. Am I overthinking it? I realize the extension is open source, of course.
My first free software search project was Searx, a privacy respecting metasearch engine, but because of the limitations of the metasearch concept, I've decided to take a different approach.
Hister builds a personal search index from pages you visit, bookmarks, browser history, local files, and crawled websites. It stores extracted content with offline result previews, so information remains searchable even when the original page changes or disappears. It supports full text and semantic search, can run entirely on your own machine, and includes a web interface, command line tools, and an MCP endpoint for assistant integrations.
Project page: https://github.com/asciimoo/hister
Tiny read-only demo: https://demo.hister.org/
Immediately interested and will check it out, thank you! I've wanted a "search stuff you've seen online" tool for a long time, but everything seems to be research-oriented or "archive but don't search" or some weird combination that means it's nigh useless to me. I've got decades of bookmarks and archives and I've kinda been stuck grepping them at best (it's rare but I do sometimes want a page I saw once three years ago and I love having that option), while hoping someone would build something better.
One question if ya don't mind, while I explore: any chance of singlefile support? Content-extraction is useful in lots of situations (e.g. wallabag) and it's a great default, but sometimes it fails and sometimes you really do want the page, relatively close to how it actually was. Singlefile does that much better than most, and it does so well enough (and manually-handle-able enough if needed) that I don't feel any desire to switch to WARCs or similar.
Though specifically I'm probably looking for something like "content-extract everything" + "key combo to save singlefile version too" + "upload singlefile archives to backfill / recover". Like 99% of the time content extraction is preferred, and I'm glad to see it... it's just not always enough, and having to go elsewhere for exceptions breaks a lot of the utility.
> One question if ya don't mind, while I explore: any chance of singlefile support?
Yes, partially. Hister can already import HTML files created by SingleFile, but there is no direct integration yet. In the longer term, I would like the SingleFile extension to be able to send snapshots directly to Hister.
Overall I really like what I'm seeing, it ticks a lot of important boxes for me and it's pleasantly straightforward. Hopefully I'll find time to contribute!
Hister always stores the original material.
> That way you could also switch from an extracted view to a "full" view in the UI.
It isn't even needed, we just need a SingleFile specific extractor (an interface in Hister to parse specific page content and provide custom previews) that provides the full original HTML for the preview panel.
> Hopefully I'll find time to contribute!
I'd appreciate it. <3
Thank you again!
Would Hister support this basic workflow? I'd love to retire my own software.
The next phase was going to move to a recording proxy.
[0] https://github.com/canvas-ui/canvas/tree/main/apps/browser-e...
https://beepb00p.xyz/promnesia.html
I'm hacking on a in-process DB on top of LMDB+Lance(for now, hilbert space kung-fu with a custom matryoshka embedding setup with separate spatial + temporal + internal and content derived anchors will replace that ~last-century~ last-year tech) + roaring bitmaps as the primary indexing engine with the same or similar goal[0] and will definitely deep-dive into yours.
I'm also trying to index users unstructured documents and workflows(tabs, emails, files, notes, identities etc)
- Organize them into semantically meaningful user or agent created context or directory-like virtual trees (the same photo of a nice kitchen may be surfaced under `/travel/barcelona` and `/arch/interieour/kitches`)
- ..where tree nodes are mapped to bitmaps - `/travel/barcelona` does a fast and cheap `travel` AND `barcelona`, want to "zoom-out" you just go one directory up to `/travel` and see all documents tagged with travel)
- You can use multiple timelines - extract that fancy md-converted en-wiki hf dataset into a wikipedia db dataset + timeline, tag your personal timeline as "personal" - wanna know the zeitgeist of your grandmothers birth date - search for it with timelines personal + wikipedia in layered mode and you'll get everything that happened or was happening during that time.
- You can have long-running stateful query sessions and refine your searches dynamically - search for "winter" and get all documents with a winter scenery or mentioning winter - refine with "nice view" then "laptop" - citing a recent example[1]
- Documents have relations that would be cumbersome to map in a virtual tree structure(worth an experiment due to the zoom-in/out you get with context bitmap trees though) - hence on top of the initial structure you can use graph edges(also powered by bitmaps - as most indexes are)
- All vector queries always run on top of a candidate set you get by the bitmap/bitmap-based filter algebra hence searching through 100k+ docs is usually pretty fast
Anyhow, let me stop here, thank you once again!
[0] https://github.com/canvas-ui/canvas-synapsd (sorry for the sloppy ai readme, no time to resurrect my old one with the updated APIs)
[1] https://demo.cnvs.ai/pub/c/aks6zaf8
One thing at first look, like a little feature request ;).
Personally I think it would be very nice if it was possible to index different sites to different search indexs. So you could separate different stuff, like job, different projects you have, and other stuff and then last everything else etc to there on search index db.
So it would be possible to have like "profiles" you could easy choose from. So in the addon, when you click on the hister icon on the toolbar there would be a list of profiles that you could choose where to index the site to. But also a setting to choose which domains and sites that always should index to one search index profils. And all sites that dont is added to a profile is added to the default index instead.
For me this would make it much more easy to find stuff and sort out what Im looking for (for me things get very "thing"/project based what I want to find), so having this..
1 - The "profiles" would become like an important filter. When you search you could easy mark one or more "profiles" which you would search from, and you remove alot of unwanted data automatically, specially when you index every site you visit by automatically.
2 - It would also make it easy if the db gets big over time to remove data that is less important later on and that you dont need to be index anymore.
3 - It would also make it easier to backup only the most important search index DB:s if the DB for the differnt search index:s where in different folders/files. I guess the "default index" that index all sites could easy get big, but there it would be alot of not important data. So would be nice to be able to easy just use a backup program and only backup the only important search databases, to save space.
Thanks
Would Hister be suitable for this? Can it index mbox files? Would it handle this amount of data? Does it have a search API so I can build an MCP server?
The indexer it uses (Bleve) can handle millions of records according to their docs, but sure that Hister would be the best choice for this task. I'd probably use Meilisearch (https://www.meilisearch.com/).
> Can it index mbox files?
Not yet.
> Does it have a search API so I can build an MCP server?
It has both search API and MCP server endpoints.
I often find myself irritated because I read an article on my phone 6 months ago and the history is gone.
Automatic page capture on mobile currently requires Firefox, since mobile Chrome does not support browser extensions.
would've been great with a more liberal license
Not in the near term. Right now I am focused on developing Hister rather than operating a hosted service. There are already plenty of centralized hosted search engines, so my longer term interest is in federation and distributed search. I want to make the core system mature first.
> would've been great with a more liberal license
It depends on how do you define liberal. =] I chose AGPLv3+ because I want Hister to remain free software and available to their users.
i still think it could've benefited by Apache 2.0 which more or less gets you to your goals
Academics usually take it to mean the Danube river instead, but in conspiracy/New Age contexts, the association with Hitler is prevalent, and mentioned in several pop culture works, including at least one feature film.
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]
[1]: https://wmchen.com/blog/revisiting-hister/
[2]: https://hister.org/datasets
What tools or features would Hister need to support your complete search workflow?
- let us say I want to index every blog ever listed on HN
- should be a small subset of the 400 billion pages out there on the internet no?
- First I need to gather data, what do you use to load so many webpages rapidly? asyncio with aiohttp in python? are there better options?
- how do you handle proxies? rotation? are there libraries you recommend for this?
- what about pages that use cloudflare? or block your request or present a captcha or a challenge of some kind?
- what are the filetypes you collect? only html or media as well?
- where and in what format do you store all these collected files? flat file storage? duckdb? postgres? hstore? something else?
- what is the frequency at which you refresh each page? once a day? once a week? something else?
- what kind of pre-processing do you use on the collected data? remove extra spaces? special characters? some kind of complex regex pipeline? LLM?
- how do you match the incoming query with processed data? simple text matching? regex? vector embedding match? something else?
For the semantic search is there any chunking/processing that happens with the content or do you need to be diligent about having a large embedding context (and/or small content)?
https://github.com/rumca-js/Internet-Places-Database
Also I maintain android app that can be used to search places.
https://f-droid.org/en/packages/io.github.rumcajs.offlineweb...
I believe hister you have to fill in with your data, right?
It also has a "public mode" where anyone can search the indexed content, but only authenticated users can add or modify it.
Also indexed data is persisted on a per-user basis, so you got this isolation and certainty that your searches will not be polluted by your family's
Planning to contribute significant improvements to the vector search side here.
Got enough Hitler in my life with Google...
I feel this is a weird spin on the "Clicks to Hitler" game
Where it says that “Hister” is a variant of its Latin name.
I think the name is lovely with no reason to change. Surely there are few people who would connect this with silly medieval superstition. Perhaps you can take part in wiping out that association.
It's a terrible choice of name
Unlike say Hipster with the descender on the p.
Webtm.io
All open source and small enough to deploy. I deploy to cf webworkers so it’s the only place it’s tested.
One cool thing is we work on iOS, chrome and friends, Firefox and pretty much everywhere. We do require you bring your own LLM though.