Provider orchestration
Route prompts across Claude, OpenAI, Gemini, Bedrock, and optional local Ollama with configurable failover and model selection.
AI chat operations platform
Route prompts across Claude, OpenAI, Gemini, Bedrock, and optional local Ollama with configurable failover and model selection.
Conversations are sessionized, summarized, and embedded for semantic retrieval. The bot remembers what was discussed across sessions, not just the last few messages.
Ship useful workflows out of the box: summaries, introductions, event planning, moderation signals, and owner digests.
Connect live data workflows for weather, transit, venues, news, books, and game utilities without custom glue code.
You can test Garbanzo right now at demo.garbanzobot.com without installing anything.
Status: Production runtime
Best for community and self-hosted group operations.
Status: Official API runtime
Best for stable channel operations with predictable webhook behavior.
Status: Official API runtime
Supports production bot interactions with strong server/channel tooling.
Status: Adapter target
Planned through the shared adapter architecture.
Run one chat runtime with multiple AI backends, practical integrations, and operational guardrails designed for active communities.
Support keeps maintenance, releases, and support response sustainable.
Garbanzo Bean is operated as an independent software project focused on reliable AI chat operations for communities.