Congratulations, this is a nice concept, this is actually usefull for large corporates with distibuted teams. The only issue is it is very hard these days to sell developer tools to enterprises, every tool any corporate is using these days are trying to inflate their offering with AI stuff. But if you can close them, definately this will be usefull. One more thing, your pricing is quite aggressive at 30/seat.
It's still probably the best tool to navigate, visualize and understand complex codebases, which is particularly important now with AI coded repos. I keep looking for alternatives but they are all notably worse.
About a month ago I spent a few frustrating hours building it from source for my system and making it work, and I've enjoyed using it as my main IDE since.
I wish I had the time to make a fork and bring in a newer version of VSCode. If anyone takes it up I might help at least.
Yeah I agree it's a shame. Unfortunately coding has changed fast and I was not confident that the editor was in the correct direction with AI coding becoming so prevalent.
I think there is a lot of value with "reconnecting" with your codebase, so I do have some plans to bring the core concept of Haystack back in one form or another.
I am very confident that editor is the direction I want coding to go. I really encourage you to consider picking it up again. Debugging complex graphs of function calls is not going away quite yet.
100% - tried forking VSCode too, scratched it. Way too much work to restyle, and I started wondering if IDEs should even be code-focused anymore. Built https://getmesa.dev instead, a canvas-based IDE for Claude Code & other CLI Agents. Kinda stuck on direction though - lmk if you've got ideas or wanna help out : )
I'd certainly love to help! I've also been working on a canvas for building agentic flows and workspaces, and I'm actually at an intersection to what you're doing with that application you're showing off. I have a simpler node based stream system for flight control of arbitrary agent actions, meant to be operated by one or more orchestrating agents. Eg, the agent understands the surface it can create nodes with, generate dynamic tools and functions within the node specs, and then write graphs of them together so it's live and observable. Extremely intersectional, haha.
Happy to chat whenever you'd like, I can stalk a contact for ya or vice versa!
Yes, I think that having the code completely checked by AI is not a good idea, but an AI that says, "Check these," because they are noteworthy. This is my idea of a future where I hope AI is like the movie "Limitless." That is, it supports you in improving yourself and giving you greater capabilities, not in replacing you entirely.
I like this idea. To be blunt, would this have more features than hooking up Claude/Gemini/Codex and saying "If - at any point - you're unsure, step back and let a human review"
1. A centralized review mechanism for a team or org that operates on coding agent conversations in addition to diffs (and the codebase). It evaluates multiple different variables (e.g. how sensitive are the changes, how much did the author do to derisk, and what did the author's coding agent gloss over) and helps enforce your team's guidelines moreso than just an individual's prompt
2. Adversarial review that operates in addition to other AI review agents (e.g. BugBot, or Greptile) and filters any comments to only the things the author cares about. This helps cut down on the "AI reviewer battleground" that is present in pull requests
3. A review interface that allows human reviewers to quickly understand what the author did to verify their changes and focus on the author's design decisions
We actually jury-rigged all of this together before building Haystack, but found that it doesn't scale to the team level (since every individual has their own ideas/opinions of what constitutes a human review).
We also found that reviewing through purely Claude Code/Codex was slow and difficult because stuff like author traces are not pre-processed and you have to get your agent to specifically explore/understand them.
I think that AI coding is only going to grow in usage. I also think that AI more-or-less writes working code that is faithful to the user's prompt.
At the same time, AI does not write code that's easy for humans to review, as it writes very verbosely and with the goal of getting the job done vs producing readable code.
I really don't think humans should be reviewing the absolute mountains of LLM-produced code; doing so would only exhaust someone's cognitive budget.
Therefore, I think that bug finding should be left to AI to review, while architectural decisions should be reviewed by humans.
Congratulations, this is a nice concept, this is actually usefull for large corporates with distibuted teams. The only issue is it is very hard these days to sell developer tools to enterprises, every tool any corporate is using these days are trying to inflate their offering with AI stuff. But if you can close them, definately this will be usefull. One more thing, your pricing is quite aggressive at 30/seat.
Man, it's such a shame that you pivoted away from the canvas-based editor concept, it was such a pleasure to use, it's so much better than tabs.
https://github.com/haystackeditor/haystack-editor
It's still probably the best tool to navigate, visualize and understand complex codebases, which is particularly important now with AI coded repos. I keep looking for alternatives but they are all notably worse.
About a month ago I spent a few frustrating hours building it from source for my system and making it work, and I've enjoyed using it as my main IDE since.
I wish I had the time to make a fork and bring in a newer version of VSCode. If anyone takes it up I might help at least.
Yeah I agree it's a shame. Unfortunately coding has changed fast and I was not confident that the editor was in the correct direction with AI coding becoming so prevalent.
I think there is a lot of value with "reconnecting" with your codebase, so I do have some plans to bring the core concept of Haystack back in one form or another.
I am very confident that editor is the direction I want coding to go. I really encourage you to consider picking it up again. Debugging complex graphs of function calls is not going away quite yet.
100% - tried forking VSCode too, scratched it. Way too much work to restyle, and I started wondering if IDEs should even be code-focused anymore. Built https://getmesa.dev instead, a canvas-based IDE for Claude Code & other CLI Agents. Kinda stuck on direction though - lmk if you've got ideas or wanna help out : )
I'd certainly love to help! I've also been working on a canvas for building agentic flows and workspaces, and I'm actually at an intersection to what you're doing with that application you're showing off. I have a simpler node based stream system for flight control of arbitrary agent actions, meant to be operated by one or more orchestrating agents. Eg, the agent understands the surface it can create nodes with, generate dynamic tools and functions within the node specs, and then write graphs of them together so it's live and observable. Extremely intersectional, haha.
Happy to chat whenever you'd like, I can stalk a contact for ya or vice versa!
Yes, I think that having the code completely checked by AI is not a good idea, but an AI that says, "Check these," because they are noteworthy. This is my idea of a future where I hope AI is like the movie "Limitless." That is, it supports you in improving yourself and giving you greater capabilities, not in replacing you entirely.
Haystack is a popular Framework for building AI agents: https://haystack.deepset.ai/
I like this idea. To be blunt, would this have more features than hooking up Claude/Gemini/Codex and saying "If - at any point - you're unsure, step back and let a human review"
I think Haystack offers:
1. A centralized review mechanism for a team or org that operates on coding agent conversations in addition to diffs (and the codebase). It evaluates multiple different variables (e.g. how sensitive are the changes, how much did the author do to derisk, and what did the author's coding agent gloss over) and helps enforce your team's guidelines moreso than just an individual's prompt
2. Adversarial review that operates in addition to other AI review agents (e.g. BugBot, or Greptile) and filters any comments to only the things the author cares about. This helps cut down on the "AI reviewer battleground" that is present in pull requests
3. A review interface that allows human reviewers to quickly understand what the author did to verify their changes and focus on the author's design decisions
We actually jury-rigged all of this together before building Haystack, but found that it doesn't scale to the team level (since every individual has their own ideas/opinions of what constitutes a human review).
We also found that reviewing through purely Claude Code/Codex was slow and difficult because stuff like author traces are not pre-processed and you have to get your agent to specifically explore/understand them.
Let AI review the code written by AI? interesting...
I think that AI coding is only going to grow in usage. I also think that AI more-or-less writes working code that is faithful to the user's prompt.
At the same time, AI does not write code that's easy for humans to review, as it writes very verbosely and with the goal of getting the job done vs producing readable code.
I really don't think humans should be reviewing the absolute mountains of LLM-produced code; doing so would only exhaust someone's cognitive budget.
Therefore, I think that bug finding should be left to AI to review, while architectural decisions should be reviewed by humans.
Almost as interesting as humans reviewing code written by a human intern or temp or an offshore dev.
On second thought, exactly as interesting.
Everyone should have been doing XP for 30 years now:
https://en.wikipedia.org/wiki/Extreme_programming
Just to say great idea. But the name "Haystack" is used by several dozen things FWIW :)