You've done the interesting work: the agent is built, deployed, and reachable at an endpoint. And then reality hits — to actually talk to it, someone has to build a chat UI. Again.
AgentChat Hub exists so you never have to. If your agent speaks A2A, paste its endpoint into the app and you're chatting in seconds. No wrapper, no middleware, no frontend repo to maintain.
What you skip
- No custom chat interface to design, build, and host
- No streaming/rendering logic — animated responses, task progress, and tool call visibility come built in
- No auth screens or user accounts to bolt on
- No app to distribute — your users download AgentChat Hub once and add any agent
How it works
AgentChat Hub reads your agent's public metadata from its endpoint, so connection is a paste-and-go flow. Add the agent, start a conversation, and watch it think: streaming tokens, task updates, and tool calls appear in real time.
Need to iterate on the agent itself? Refresh the agent in the app and it re-reads the latest metadata. Your conversation history stays on the device, so testing a redeployed agent doesn't lose your context.
When this is the right tool
If you're building agents for yourself, for a team, or for clients — and the interface is a chore, not the product — AgentChat Hub is the missing piece. Build the agent. Skip the frontend. Start chatting.
Try it yourself
The interface your agent was missing.
Free on the App Store. No account, no tracking — your conversations stay on your device.