[GH-ISSUE #274] Connectome — open-source AI.OS that goes beyond chat UI to agent-generated dynamic interfaces #234

New Issue

2026-06-05T16:36:23-04:00

yindo commented

2026-06-05 16:36:23 -04:00

Originally created by @AvielCarlos on GitHub (Apr 27, 2026).
Original GitHub issue: https://github.com/langchain-ai/agent-chat-ui/issues/274

Hey LangChain team 👋

agent-chat-ui is a clean reference implementation — a web app for interacting with any LangGraph agent via chat. It's a natural starting point for agent UX.

I'm building Connectome and I wanted to share a different direction we're exploring with agent interfaces — one that might be interesting to the LangChain community.

Rather than agents that respond to chat messages, Connectome's Ora generates personalized JSON screen specs in real-time that drive the entire UI of a React Native app. The interface isn't a chat window — it's a dynamic environment constructed by 10+ agents who know who you are and what you need right now.

It's the next step beyond agent-chat-ui: instead of asking "what should I show the user when they send a message?", we ask "what should the user's entire experience look like right now, without them asking?"

Stack: FastAPI + PostgreSQL + pgvector + Redis + Expo React Native

This is early but the architecture is working. We think there's a lot to learn from how LangGraph handles state and from what agent-chat-ui has figured out about agent UX.

We're a DAO. Contributions earn CP tracked by an LTV AI. First 10 to reach 3,000 CP become Founding Stewards — permanent governance rights.

Worth talking about if any of the LangChain team is interested in human-flourishing applications of agent technology.

👾 https://t.me/ascensioncommunity
🌐 https://avielcarlos.github.io/ascension-site

— Avi / Aviel Carlos | @neoandromeda | carlosandromeda8@gmail.com

Originally created by @AvielCarlos on GitHub (Apr 27, 2026). Original GitHub issue: https://github.com/langchain-ai/agent-chat-ui/issues/274 Hey LangChain team 👋 agent-chat-ui is a clean reference implementation — a web app for interacting with any LangGraph agent via chat. It's a natural starting point for agent UX. I'm building **Connectome** and I wanted to share a different direction we're exploring with agent interfaces — one that might be interesting to the LangChain community. Rather than agents that respond to chat messages, Connectome's **Ora** generates *personalized JSON screen specs* in real-time that drive the entire UI of a React Native app. The interface isn't a chat window — it's a dynamic environment constructed by 10+ agents who know who you are and what you need right now. It's the next step beyond agent-chat-ui: instead of asking "what should I show the user when they send a message?", we ask "what should the user's entire experience look like right now, without them asking?" **Stack:** FastAPI + PostgreSQL + pgvector + Redis + Expo React Native This is early but the architecture is working. We think there's a lot to learn from how LangGraph handles state and from what agent-chat-ui has figured out about agent UX. **We're a DAO.** Contributions earn CP tracked by an LTV AI. **First 10 to reach 3,000 CP become Founding Stewards** — permanent governance rights. Worth talking about if any of the LangChain team is interested in human-flourishing applications of agent technology. 👾 https://t.me/ascensioncommunity 🌐 https://avielcarlos.github.io/ascension-site — Avi / Aviel Carlos | @neoandromeda | carlosandromeda8@gmail.com

yindo referenced this issue

2026-06-05 16:37:02 -04:00

[PR #234] [CLOSED] Staging #240

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: langchain-ai/agent-chat-ui#234