I started building my own version of this in Superlogical Rex.
It's a bit hard to decide where to hook in.
Rex has events, harnesses have hooks, terminal output can set program status, git has hooks.
I'm not sure I actually want every session from every machine saved. That's a lot of noise especially when half of coding sessions are experimental / throwaway.
But at the same time its not that much data that's easy to ripgrep / LLM through.
"Half of coding sessions being throwaway" is exactly why I stopped keeping them per session. Mine hang off a work item, and each item has its own folder with its session logs.
I basically throw them away and also record their summary in a vector DB to be able to find why I did something, if I'm ever curious about it!
Otherwise I'm gonna need a tool to help me manage my sessions :D
I suspect this is one of those problems "everyone" has built their own solution for. Including me!
I built slivingdoc[0] with an alternative approach: instead of syncing exact sessions (which are often quite noisy, and not so useful), the agents gets a living document with automatic merge conflict resolution backed by s3 ("infinite" storage). Idea being that the agents are prompted to summarize their context into slivingdoc as normal utf-8 files. So this way only you only need to add the stdio MCP server to all your harnesses which supports it and it just works!
Some harnesses are _super_ hyped about this concept. Codex needs no prompting at all to start communicating. I got the idea from enabling agentic forums via all the openAI hacks.
Really enjoyed Atuin for a few weeks, until I realized it doesn't have correct `EFSRO` (Error: Read-Only File System) handling. So when my `/home` mount had a raid array drive failure it broke my local shell. Issue has been sitting open for almost a year. Pretty disappointing.
"Just" filing an issue takes time and effort. Providing a patch takes significantly more time and effort. Potentially the difference of a few minutes vs a few days, weeks, or longer, depending on the issue at hand.
There is value in filing well written issues, and no one should be shamed for "just" filing an issue to the people best equipped to resolve it, rather than trying to fix it themselves.
You should probably have read the comment that person was responding to, because you seem to have taken the response out of context. OP did file an issue over a year ago, and didn't submit a patch, so they did indeed "just" file an issue. Then they proceeded to whine about no one working on their issue for free.
Didn't read as whining to me, just accurately describing a bug with an anecdote about how it negatively affected them. Then a developer turned up as if hurled from the sky and fixed it! The system works!
I've been maintaining a tool that does similar: https://github.com/emberian/cv. High performance parsers of the major harness formats at several versions. Not as much autosyncing. Great results using custom prunings compared to compaction summarization.
I love Atuin, but have some reservations about a shared server keeping all my shell history and agent sessions. Am I being too paranoid and missing out on convenience?
Same, I really wanted to share history, but when I realised I was going to need to send things to a remote server I instead wrote a tiny script to search through backups that are shared across machines by Syncthing.
I just search over $BACKUPS/*/latest/local/share/fish/fish_history with grep and skim
I've built this for myself, but not across multiple machines.
It scans the session logs of all coding harnesses I use, one extra thing it does is create embeddings to allow hybrid BM25/semantic search.
I think these kinds of tools are on of the few times I've spent time "using AI to work on my AI tools" that was actually a good use of time.
It's so useful to be able to say "search sessions for the time I was working on XYZ 3 months ago"
I've been experimenting with the same thing! Still don't have enough evidence to be able to say if it's beneficial or not, but hopefully will have a conclusion soon
I've gone a different route and prefer using the native harnesses (Claude, Codex), but with multiple accounts due to the limits.
In order to easily switch between accounts I've build Switcheroo, which I've been using for the past 2-3 months now. I've decided to open source it: https://github.com/FTCHD/switcheroo
Works with other AI CLIs like Gemini, and other various CLIs like Vercel (if you're managing multiple accounts like me), Netlify, Github etc
Interesting, does it avoid the little song and dance to switch between claude accounts? I think ideally there could be a coordinator session that monitor session usage and tell other sessions to pause their work at around 95%, and run your tool to switch those sessions to another account. Then tell the other sessions to resume their work. Ideally that session could stop and start the claude background processes, which avoids remote control disconnection when switching session. And that session would of course be isolated to avoid being affected by the account change
It's not clear which parts of Claude state are safe to sync and which you want to keep local to one machine. I'm doing something with syncthing trying to keep agent session history available and little bits of it keep breaking. I haven't looked at this new feature in atuin does it have a really good understanding of what is worth syncing?
One of the Atuin engineers here. Not sure wht your definition of safety is, but Atuin doesn't attempt anything special in terms of deciding what to sync. Pretty much everything is synced.
All your sessions are end-to-end encrypted so only your machines have access to your sessions! Happy to answer any other questions!
https://github.com/comalice/pi-tools contains a 'memory everywhere' module that makes all sessions BM25 vectors, and surfaces a set of memory tools to enable agents to search. You can add a new machine to the mix by registering a 'spoke'. The app runs on the 'hub'. I want to containerize the core app soon.
You can also change which computer is the hub fairly easily.
Will this work for Claude code web sessions?? I am DYING to be able to control and sync between them locally. IE take advantage of anthropics compute to build my toolchain which, when running many sessions locally uses significant ram.
I love this new AI era of coding where the HN submission is about a new way to manage agent, and almost literally all the comments are "I just finished implementing this myself"
It's almost like a parody.
This is something I've noticed more widely and is something I think we're going to have to grapple with sooner or later. If anyone can make any software, what happens to our shared computing reality? Now one can steal features from 3 or 4 similar projects and compose exactly what I want, and there are now 5 similar projects.
Similarly, I've already had many issues with project shunning me because I come with AI assisted code, no matter the quality of the contribution. With NixOS it's extremely simple to carry patches, and I do this with a number of projects (k3s, kata, Ardour).
It feels like long-term forks are something we're going to have to accept as the new reality, and find some way to weave together in a new way, rather than as just an escape hatch to route around maintainer issues.
I am not sure that question really applies here with AI tooling. This is still an immature field where people are discovering optimal workflows.
Not to sound too Dropboxy, but you could get an 80% solution to this problem with rsync and a hundred lines of python. Or just punt it all into git. I regularly see people inventing their own workflow utilities. Less so someone vibing a custom IDE.
Yes i too faced the problem, lost valuable content across group of chat and not remembering in which chat it was. Tried searching but couldn't find it. Even asked ai itself does about the content whether it remembered.
Correct me if I am wrong, but I imagine this would hammer one's token usage to read past entries to resemble context across machines/terminals and models?
I don't think this, by default, has any effect on anybody's token usage, it's just a way to save these sessions if you need to reference them again in the future. Unless of course you choose to provide a previous session to a new agent, but that should be expected.
Probably a little late to this, because herdr has really eaten this segment. With their new Mac app, I don't see myself going back to atuin for this kind of experience. In fact, I get a little peeved using it on a vanilla terminal.
What peeves me is the poor responsiveness when navigating history. Pressing up or ctrl+r has this lag which throws off my typing. I'm on a pretty well specced setup, too. I used to be able to tolerate it, but since I barely do things directly from the command line anymore and have Claude worry about most of the command construction, the experience can be particularly grating.
Its multi host mode is pretty limited still. Like especially with a web ui, it will just install a random, auto-updating binary on the other machine, which I didn't feel comfortable with. I would much rather have herdr itself manage the connection completely, not a plugin.
I've spent the last couple of months putting a lot of time and energy into isolating my agents from each other: aggressive sandboxing, regular scanning of transcripts for agent-leaked credentials, per-agent encrypted credential stores, even physical (and network) separation of machines with different areas of work on them. Needless to say, as much as I have found atuin an excellent manager of my command-line history, I will be avoiding this.
Yes and the bigger problem for me is the agent can write to that file too. The file that decides what it is allowed to do is just another file in the repo for it. I now treat the agent config files as something to protect from the agent itself, not only from other people.
It’s likely because you can vibe code the bones of one in a few hours.
This week I got Claude to build me a browser-based session manager to keep track of sessions and history of tasks. Works well enough for me and I’ll add features as I need.
I started building my own version of this in Superlogical Rex.
It's a bit hard to decide where to hook in.
Rex has events, harnesses have hooks, terminal output can set program status, git has hooks.
I'm not sure I actually want every session from every machine saved. That's a lot of noise especially when half of coding sessions are experimental / throwaway.
But at the same time its not that much data that's easy to ripgrep / LLM through.
Nice! Rex is super cool. We have an Atuin Threads integration for Herdr already, definitely exploring building one for Rex too
We actually don’t use hooks at all at the moment, and instead rely on parsing session files
Agents are definitely super good at searching and finding relevant data amongst noise!
"Half of coding sessions being throwaway" is exactly why I stopped keeping them per session. Mine hang off a work item, and each item has its own folder with its session logs. I basically throw them away and also record their summary in a vector DB to be able to find why I did something, if I'm ever curious about it! Otherwise I'm gonna need a tool to help me manage my sessions :D
Unrelated, but if you have an invite to Superlogical Rex, would appreciate one. gajus@gajus.com
Honestly, just go sign up on the Superlogical website, the queue is short, you should get an invite quickly enough.
As far as I know(?) beta testers don't have the ability to give out invites anyways. At least I don't.
Interesting. I signed up a while ago and did not receive an invite.
Anything you can share yet? I was about to point Claude at the same thing in Rex and would rather not burn the tokens to come to the same result
I suspect this is one of those problems "everyone" has built their own solution for. Including me!
I built slivingdoc[0] with an alternative approach: instead of syncing exact sessions (which are often quite noisy, and not so useful), the agents gets a living document with automatic merge conflict resolution backed by s3 ("infinite" storage). Idea being that the agents are prompted to summarize their context into slivingdoc as normal utf-8 files. So this way only you only need to add the stdio MCP server to all your harnesses which supports it and it just works!
Some harnesses are _super_ hyped about this concept. Codex needs no prompting at all to start communicating. I got the idea from enabling agentic forums via all the openAI hacks.
[0]: slivingdoc.dev
Similar to one I designed (but haven't built yet) that uses the agent itself to resolve conflicts: https://github.com/cretz/theparlor/blob/initial-work/doc/des...
Really enjoyed Atuin for a few weeks, until I realized it doesn't have correct `EFSRO` (Error: Read-Only File System) handling. So when my `/home` mount had a raid array drive failure it broke my local shell. Issue has been sitting open for almost a year. Pretty disappointing.
Did you provide a patch or just filed an issue?
"Just" filing an issue takes time and effort. Providing a patch takes significantly more time and effort. Potentially the difference of a few minutes vs a few days, weeks, or longer, depending on the issue at hand.
There is value in filing well written issues, and no one should be shamed for "just" filing an issue to the people best equipped to resolve it, rather than trying to fix it themselves.
You should probably have read the comment that person was responding to, because you seem to have taken the response out of context. OP did file an issue over a year ago, and didn't submit a patch, so they did indeed "just" file an issue. Then they proceeded to whine about no one working on their issue for free.
Didn't read as whining to me, just accurately describing a bug with an anecdote about how it negatively affected them. Then a developer turned up as if hurled from the sky and fixed it! The system works!
Very disappointing when something you get for free isn't the way you want it to be.
Atuin engineer here. I assume you're referring to https://github.com/atuinsh/atuin/issues/2999.
Will fix.
Fix in-flight: https://github.com/atuinsh/atuin/pull/4711/changes
Thanks! I've hit this issue before also.
fixed in the next release :)
I've been maintaining a tool that does similar: https://github.com/emberian/cv. High performance parsers of the major harness formats at several versions. Not as much autosyncing. Great results using custom prunings compared to compaction summarization.
I love Atuin, but have some reservations about a shared server keeping all my shell history and agent sessions. Am I being too paranoid and missing out on convenience?
Same, I really wanted to share history, but when I realised I was going to need to send things to a remote server I instead wrote a tiny script to search through backups that are shared across machines by Syncthing.
I just search over $BACKUPS/*/latest/local/share/fish/fish_history with grep and skim
Nice solution
Atuin offers a self-hosted backend server for syncing, so you don't need to trust anybody
It's centralized, which is annoying unless you can trust A machine you own to be up and reachable by all others all of the time.
Atuin is open source, you can just host your own server?
Fair enough, I missed that. Will give it a try!
It is encrypted client side. Or you can run your own server.
> Am I being too paranoid and missing out on convenience?
No, I'd say it is more-so a lack of reading the article that is causing you to miss out.
I did read the article but missed that. Paranoia can definitely co-occur with distracted reading, they're independent dimensions ;)
Thanks for the info, will give it a try
I cant get used to it taking over the entire screen when scrolling history
I faced this too! Turned out it’s an easy config change to limit to height of your choice
atuin config set inline_height 5
export ATUIN_NO_MOUSE=true
But… you don't have to sync anything?
I've a love affair with Atuin but this may turn into a full blown obsession
love to hear it! lmk if there's anything we can do to be better (ellie@atuin.sh)
It’s funny I use Atuin multiple times a day without even realizing it. Going to try this out they put out quality software.
I've built this for myself, but not across multiple machines. It scans the session logs of all coding harnesses I use, one extra thing it does is create embeddings to allow hybrid BM25/semantic search.
I think these kinds of tools are on of the few times I've spent time "using AI to work on my AI tools" that was actually a good use of time.
It's so useful to be able to say "search sessions for the time I was working on XYZ 3 months ago"
I've done exactly this.
I'm also experimenting with a 'dream mode' that extracts facts, and a 'librarian' that stores those facts in the MD wiki.
I don't know if there's a benefit to that yet, but I'm going to find out.
I've been experimenting with the same thing! Still don't have enough evidence to be able to say if it's beneficial or not, but hopefully will have a conclusion soon
posted on X a couple a days ago how atuin threads saved me from loosing the code I was working on and I accidentally removed uncommited worktree https://x.com/nklmilojevic/status/2118401844401541604
super cool use case! still exploring if there's anything fun we can do with source control
My version is very unixy:
$ ctools resume ssh://laptop.local/code/project
Goes to laptop.local, looks for sessions done in that path, finds the most recent, copies it locally and resumes it.
There's also ls, grep, copy, cat... It looks and feels like the existing tools, that's the point.
Workflow is the thing that matters here and I'm terminal first
I'll make it more flashy soon https://github.com/day50-dev/ctools
I've gone a different route and prefer using the native harnesses (Claude, Codex), but with multiple accounts due to the limits.
In order to easily switch between accounts I've build Switcheroo, which I've been using for the past 2-3 months now. I've decided to open source it: https://github.com/FTCHD/switcheroo
Works with other AI CLIs like Gemini, and other various CLIs like Vercel (if you're managing multiple accounts like me), Netlify, Github etc
this works with native harnesses, and is neither a harness nor a wrapper of them
atuin stores, searches and syncs agent sessions regardless of harness or setup
Interesting, does it avoid the little song and dance to switch between claude accounts? I think ideally there could be a coordinator session that monitor session usage and tell other sessions to pause their work at around 95%, and run your tool to switch those sessions to another account. Then tell the other sessions to resume their work. Ideally that session could stop and start the claude background processes, which avoids remote control disconnection when switching session. And that session would of course be isolated to avoid being affected by the account change
It's not clear which parts of Claude state are safe to sync and which you want to keep local to one machine. I'm doing something with syncthing trying to keep agent session history available and little bits of it keep breaking. I haven't looked at this new feature in atuin does it have a really good understanding of what is worth syncing?
One of the Atuin engineers here. Not sure wht your definition of safety is, but Atuin doesn't attempt anything special in terms of deciding what to sync. Pretty much everything is synced.
All your sessions are end-to-end encrypted so only your machines have access to your sessions! Happy to answer any other questions!
Fascinating. I just finished implementing this.
https://github.com/comalice/pi-tools contains a 'memory everywhere' module that makes all sessions BM25 vectors, and surfaces a set of memory tools to enable agents to search. You can add a new machine to the mix by registering a 'spoke'. The app runs on the 'hub'. I want to containerize the core app soon.
You can also change which computer is the hub fairly easily.
Will this work for Claude code web sessions?? I am DYING to be able to control and sync between them locally. IE take advantage of anthropics compute to build my toolchain which, when running many sessions locally uses significant ram.
not yet but we're definitely interested in figuring out how to make that happen!
I've been using (and supporting) atuin for a while, and I find it a required bit of infrastructure. This is a nice upgrade!
Atuin is so invaluable to me, I self host an atuin server but even client only is amazing.
I think Jetbrains Air is trying to do this but not across machines.
I love this new AI era of coding where the HN submission is about a new way to manage agent, and almost literally all the comments are "I just finished implementing this myself" It's almost like a parody.
This is something I've noticed more widely and is something I think we're going to have to grapple with sooner or later. If anyone can make any software, what happens to our shared computing reality? Now one can steal features from 3 or 4 similar projects and compose exactly what I want, and there are now 5 similar projects.
Similarly, I've already had many issues with project shunning me because I come with AI assisted code, no matter the quality of the contribution. With NixOS it's extremely simple to carry patches, and I do this with a number of projects (k3s, kata, Ardour).
It feels like long-term forks are something we're going to have to accept as the new reality, and find some way to weave together in a new way, rather than as just an escape hatch to route around maintainer issues.
I am not sure that question really applies here with AI tooling. This is still an immature field where people are discovering optimal workflows.
Not to sound too Dropboxy, but you could get an 80% solution to this problem with rsync and a hundred lines of python. Or just punt it all into git. I regularly see people inventing their own workflow utilities. Less so someone vibing a custom IDE.
Yes i too faced the problem, lost valuable content across group of chat and not remembering in which chat it was. Tried searching but couldn't find it. Even asked ai itself does about the content whether it remembered.
atuin threads solves that problem!
I’ve been a supporter of Atuin and forever, and I really like her model here.
Excited to try it!
Built this for same purpose: https://github.com/codealong/codealong
More of a desktop flavor though
Correct me if I am wrong, but I imagine this would hammer one's token usage to read past entries to resemble context across machines/terminals and models?
I don't think this, by default, has any effect on anybody's token usage, it's just a way to save these sessions if you need to reference them again in the future. Unless of course you choose to provide a previous session to a new agent, but that should be expected.
Would be great if you could disable sync for Threads but keep history sync going.
Aren't session also deleted after some time? It would be nice to keep them.
they are! claude deletes them after 30 days, but Atuin Threads keeps them synced and searchable forever :)
Probably a little late to this, because herdr has really eaten this segment. With their new Mac app, I don't see myself going back to atuin for this kind of experience. In fact, I get a little peeved using it on a vanilla terminal.
Threads doesn’t compete with herdr, the two actually go together really nicely
Atuin indexes all your past sessions, herdr multiplexes all your present sessions.
We made a mod for herdr desktop to integrate the two: https://herdr.dev/mods/atuinsh/herdr-atuin-threads/
What peeves you on the vanilla terminal?
Cool, I'll take a look at this.
What peeves me is the poor responsiveness when navigating history. Pressing up or ctrl+r has this lag which throws off my typing. I'm on a pretty well specced setup, too. I used to be able to tolerate it, but since I barely do things directly from the command line anymore and have Claude worry about most of the command construction, the experience can be particularly grating.
Chalk it up to ever more instant gratification :)
great! please let me know if you have any issues
Ahh. We've done a lot of work to improve this lately, if you're not using the daemon please try
``` atuin config enable daemon ```
and it should perform much better (it's the default on new installs)
I'm using atuin as part of my Herdr setup. Atuin is not a replacement shell or multiplexer, it's basically a really advanced shell history.
Its multi host mode is pretty limited still. Like especially with a web ui, it will just install a random, auto-updating binary on the other machine, which I didn't feel comfortable with. I would much rather have herdr itself manage the connection completely, not a plugin.
I've spent the last couple of months putting a lot of time and energy into isolating my agents from each other: aggressive sandboxing, regular scanning of transcripts for agent-leaked credentials, per-agent encrypted credential stores, even physical (and network) separation of machines with different areas of work on them. Needless to say, as much as I have found atuin an excellent manager of my command-line history, I will be avoiding this.
yes agent leak secrets “Yes, and don’t ask again” in Claude Code would also save secrets in .claude/settings.local.json.
Yes and the bigger problem for me is the agent can write to that file too. The file that decides what it is allowed to do is just another file in the repo for it. I now treat the agent config files as something to protect from the agent itself, not only from other people.
So now everyone and their mother are putting out AI harnesses/session managers aka browser tabs for AI.
It’s likely because you can vibe code the bones of one in a few hours.
This week I got Claude to build me a browser-based session manager to keep track of sessions and history of tasks. Works well enough for me and I’ll add features as I need.
they are, but this isn't one of those
atuin threads stores every turn for every harness you use, syncs it across machines, and makes it searchable and resumable anywhere.
it's not a replacement for your harness, or your harness wrapper.
Yeah this is cross browser history, not tabs.
How much electricity wasted?
idk it's actually super low on compute, doesn't use tokens, mostly just needs storage