CoChat – Group chats with multi-model AI, built on OpenWebUI
Hi HN, We're launching CoChat, which extends OpenWebUI with group chat, model switching, and side-by-side comparison. What makes it different: CoChat is designed for teams working with AI. - Group chat with AI facilitation. Multiple users collaborate in the same thread. The AI detects group discussions, tracks participants, and facilitates rather than dictates. - Switch and compare models. Run GPT, Claude, Mistral, Llama, and others side-by-side or switch mid-conversation. - Intelligent web search. Context-aware search activates only for real-time information. - Artifacts and tool calls.…
What it does
In the maker’s words, at launch
Hi HN, We're launching CoChat, which extends OpenWebUI with group chat, model switching, and side-by-side comparison. What makes it different: CoChat is designed for teams working with AI. - Group chat with AI facilitation. Multiple users collaborate in the same thread. The AI detects group discussions, tracks participants, and facilitates rather than dictates. - Switch and compare models. Run GPT, Claude, Mistral, Llama, and others side-by-side or switch mid-conversation. - Intelligent web search. Context-aware search activates only for real-time information. - Artifacts and tool calls. Generate documents and code inline. MCP tool integration coming soon. - No subscription fee. Pay for usage/tokens only at exact list price. Non-obvious things we learned building this. Through building CoChat, we've learned some surprising things about LLM behavior. I'll share two (happy to discuss more in comments). First, models don't understand they're not the only AI in the room. When you tag a new model into a conversation and ask "what do you think of Claude's response above?", the model assumes it wrote that previous response. It will defend it, build on it, or awkwardly try to reconcile the question with its false memory of writing it. We solved this by injecting model attribution into the conversation context - explicitly marking which model generated each response. Once models understand they're looking at another model's output, they engage critically rather than defensively. The quality of cross-model analysis improved dramatically. Second, LLMs have a compulsive need to "solve" group conversations. In a multi-user thread, the AI wants to answer every question and resolve every disagreement, even when humans are working something out themselves. System prompts telling it to "facilitate, don't dictate" weren't enough. We had to restructure how we frame the AI's role in group context: it's a participant who speaks when addressed, not an omniscient moderator. Getting this balance right is still ongoing - we're curious how others have approached this. We also ran into interesting challenges around memory and tool execution in multi-user contexts (whose preferences apply? whose tools get executed?) but that's probably a separate post. Why this matters: Different models excel at different tasks. Current tools lock you into a single vendor. CoChat lets you choose the best model for each task while enabling real team collaboration. We're planning to submit all updates back to the core project or maintain an active open-source fork. Try it at: https://cochat.ai Would love feedback from teams already using AI collaboratively, or anyone interested in model comparison workflows.
Does the same job
all alternatives →
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, December 2025
the whole month →- GP
Life & fun · Dec 2025 · dosaygo-studio.github.io
- JG
Hi everyone! My name's Luke and I made the original Jmail here alongside Riley Walz. We had a ton of friends collaborate on building out more of the app suite last night in lieue of DOJ's "Epstein files" release. Please AMA!
Life & fun · Dec 2025 · jmail.world
PlanEat AI▲733AI turns your health goals into a 7-day menu & grocery list
AI · Dec 2025 · planeatai.com
- 2G
Community, All the HN belong to you. This is an archive of hacker news that fits in your browser. When I made HN Made of Primes I realized I could probably do this offline sqlite/wasm thing with the whole GBs of archive. The whole dataset. So I tried it, and this is it. Have Hacker News on your device. Go to this repo (https://github.com/DOSAYGO-STUDIO/HackerBook): you can download it. Big Query -> ETL -> npx serve docs - that's it. 20 years of HN arguments and beauty, can be yours forever. So they'll never die. Ever. It's the unkillable static archive of HN and it's…
Dev tools · Dec 2025 · hackerbook.dosaygo.com
