Building open-source dev environment for agents.

Lviv, Ukraine
OpenChamber ✨ I started this project because I fell in love with @opencode as a TUI coding agent, but realized I personally didn’t enjoy living in the terminal all day. I wanted easier remote access, something that worked nicely from a phone/tablet, and a UI that felt more natural to use day-to-day. What started as a fan-made UI slowly grew into something much bigger. OpenCode now has its own web UI and desktop app, but I still keep building OpenChamber because the project naturally grew into its own thing over time. I like experimenting with workflows, tooling, and UX ideas around agents - and people seem to enjoy using it too 🙂 github.com/btriapitsyn/openc…
10
3
59
6,983
I expected the OpenCode 2 migration to be a bug hunt. It was mostly us adopting our own code. The few things we did hit, @opencode team fixed right away. @thdxr offered to help before we even asked. Thank you ❤️
3
4
152
8,616
OpenChamber 2.0 is out, on OpenCode v2. Skills, agents, MCP servers and plugins now apply the moment you save them. No restarts. My restart button is unemployed. Plus Code Mode, web search cards and a stats page. openchamber.dev/blog/opencod…
26
21
383
110,935
Hey builders 👋 Let's turn this into one giant show & tell. What are you building? Drop your site or GitHub repo 👇 Made with OpenChamber or not, doesn't matter. Find your first (or next 100) users. Get honest feedback. Get unstuck on stack or architecture. Copy this so people actually click: 🔗 link ⚡ what it does, in one line 🧰 stack 🙋 what you need: users, feedback, or advice Questions welcome: stack, architecture, deploy, launch, anything. Not an OpenChamber Q&A, just builders talking shop. We read every reply and answer where we can. One ask: drop yours, then check out 2 other projects and leave them a real comment. Give feedback, get feedback 🤝
6
2
19
1,564
By the way, if you missed yesterday's releases from @OpenAI & @AnthropicAI, here are 2 screenshots, that says it all: Benchmarks 👇👇
2
1
14
2,147
Hey guys, We have been following @typesafeai's Jev release and hype since day 1. So we were collecting data on X posts about Jev since then because we are really curious about its use cases and possibilities And so today we are sharing it publicly with you: jev.openchamber.dev/ - 5,383 pre-selected posts - from 4,160 people - 746 with code - semantic/vector search, of course - categories, etc. - updated every 4 hours Total amount of tweets processed so far: 107,038 Feel free to explore! PS - Jev is live in OpenChamber as well.
11
19
190
29,445
In the light of recent events 🤠 Thoughts?
8
2
45
2,266
By the way: the last two releases were mostly about one thing – speed, so we added a few tweaks to handle it: v1.24.1 — big session lists open faster and use less memory, smoother scrolling, and lower CPU while the agent thinks and streams. v1.24.2 — startup no longer boots MCP servers and background work for every saved project and worktree. SSH reconnect reuses the running remote server instead of piling up duplicates. Large repos stop spawning dozens of git processes. Same app, a lot less machine.
3
1
27
1,929
Testing two things in OpenChamber 1.24.1. Both off by default. → Auto model routing: pick "Auto" and each prompt gets its own model, thinking level and agent — cheap and fast for a quick question, strong for a big refactor. Rules are editable in Settings. → Safety net: in auto-accept sessions, risky actions stay on screen with a reason instead of running silently. Both run on Jev, @typesafeai decision model, so you need a Jev key (docs.typesafe.ai). To try it: set OPENCHAMBER_ROUTING_ENABLE=1, restart, and a Routing page shows up in Settings. No key or no variable — nothing changes for you. If you do have access, I'd really like to hear where it routed wrong or where the safety net got in the way. Have fun 🙂
2
3
16
1,685
In the meantime we read 8,840 tweets about @typesafeai's Jev from its first three days. Here is what we found: Speed-up: claimed 193.6×, measured median 5× Cost: claimed 444.6×, measured median 32× Latency: claimed 70–500ms, measured median 270ms Window: 15–17 Sep 2026 Link in the comments
4
12
1,352
and 10 newly born GitHub repos on/about Jev listed as well
1
146
Replying to @typesafeai
Cool use cases and numbers included in the article, of course
1
109
10 000 stars on GitHub. Give our team a round of applause, please!
9
7
127
8,495
OpenChamber.dev retweeted
Openchamber feels so smooth and polished, it’s unbelievable! I’m definitely making the switch from Codex desktop app to Openchamber.
OpenChamber ✨ I started this project because I fell in love with @opencode as a TUI coding agent, but realized I personally didn’t enjoy living in the terminal all day. I wanted easier remote access, something that worked nicely from a phone/tablet, and a UI that felt more natural to use day-to-day. What started as a fan-made UI slowly grew into something much bigger. OpenCode now has its own web UI and desktop app, but I still keep building OpenChamber because the project naturally grew into its own thing over time. I like experimenting with workflows, tooling, and UX ideas around agents - and people seem to enjoy using it too 🙂 github.com/btriapitsyn/openc…
1
4
322
Set up OpenChamber on an @exedev box this week and honestly this is how I want to work now! I don't need to expose it or use third-party tunnels. It dials out to the relay, I just scan a QR code from the app, and that's it. Their own models (Shelly, part of the subscription) just work as a provider out of the box, no extra keys or signups. Usage for their provider is easy to setup in the app. I close the laptop with agent keeps working on the box.
8
3
18
2,118
Setup on a fresh @exedev box, three commands: curl -fsSL bun.sh/install | bash && source ~/.bashrc bun add -g opencode-ai @openchamber/web openchamber connect-url --relay --qr The last one starts the server in the background, brings the relay up, and prints a QR code. Scan it from the app and you're in. Optional: "openchamber startup enable" so it comes back after a reboot.
1
5
691
Haven't posted in a while. Here's what landed in the meantime. The one I use every day: /btw – forks a temp session from your current one, inherits the whole context, answers in a panel above the composer, leaves your chat clean. Keep it or discard it.
13
2
54
3,950
OpenChamber.dev retweeted
OpenChamber又发布了新的版本。感觉作者真的在很用心的维护。这可能是我见到的,在开源项目中最用心做前端的项目了。 目前我一直在用的OpenChamber+OpenCode+Deepseek V4 Flash真的狠狠的爽到我了。
最近一直在折腾Pi agent、OMP、Zcode、Codex、Opencode 这些Agent Harness,我终于理解了我自己真正的需求是什么。 第一:我需要一个成熟稳定的GUI页面,用惯了GUI,实在是用不下去TUI,在黑框框里聊天太难受了,你要记很多的快捷键,很多的指令,才能真正的用得下去。但是GUI只要直观的点点图标、点点按钮就行了。所以我必须要一个成熟的、稳定的、美观的,GUI的页面。 所以我排除了Pi Agent 、OMP这样的产品,他们的配套的社区GUI做的实在是太差了,我自己手搓又会花很大的精力,而且它的可定制化太强,你手搓的GUI产品没有那么多的通用性。 第二:我需要非供应商锁定或者非供应商偏袒或者锁定的一个软件产品,Codex确实开放了第三方的模型配置接口,但是Codex它配置起来真的很麻烦,第三方的模型总是二等公民,这让我很不爽。Zcode它没有办法通过Oauth登录各家的coding plan,比如Kimi For Code,比如Grok Build,是没有办法在Zcode里使用的(邪修折腾除外) 所以我排除了ZCode、 Codex这种有模型供应商偏袒的软件。 第三,我需要基于我自己的需求,可以做二次开发和定制化,便于给我的客户更方便的开箱即用。所以它一定要是开源的,许可证友好的产品,方便我自己爽,也方便我的客户,用起来简单方便,不搞那么多黑科技去hack。 所以到最后,我真正能选的就是Opencode,那么OpenCode它只是一个Agent的内核,如果要给它真正选一个成熟的、稳定的、好用的GUI产品,我终于发现了真正的版本答案:openchamber 开源地址:github.com/openchamber/openc…
11
6
46
10,919