Dashboard FluxyChat Docs

How-to guides create-fluxy-chat CLI Multi-Platform Adapter Pattern Unified Chat API Troubleshooting integration Testing integrations Authentication (JWT)Migrate from Pusher / Ably / Sendbird Feature parity checklist Self-host one command Loop control StackBlitz examples fluxy.config.ts

AI agents Agent memory Agent lifecycle callbacks Tool approval (HMAC)AI agents and human handoff (HITL)Agent queue

MCP Client Integration Room as MCP server Role-scoped message visibility Hackathon cookbooks Presence-aware escalation Cross-org agent negotiation Agent room broadcast Critical action quorum External event ingest Multi-agent debate Interview eval template Presence knock template IoT telemetry starter Audience score template Audience score rollup Asymmetric session profiles Community decision simulation Live knowledge graph Room memory for operators HITL approval chain Fleet dispatch template Room OS positioning Built with patterns

Announcement channels Telephony to agent handoff

AI Agents

AI agents

In-room agents, memory, tools, and human handoff.

Agents live on the Worker. The browser talks to a room; @assistant (or your bot) replies there.

Handoff
Memory
Lifecycle
Tool approval (HMAC)
Agent queue

LLM keys (GROQ_API_KEY / OPENAI_API_KEY) stay in Worker secrets. Groq model: openai/gpt-oss-20b.

Security & auth

JWT rooms, SAML, and operator identity.

Agent memory

Provider-agnostic memory store for agents — in-memory, D1, MongoDB-style adapters.

Ask AI

FluxyChat documentation

Ask anything about FluxyChat — SDK, Worker API, agents, or deployment.

Answers are generated from docs context. Verify critical details in the guides.