Privacy-preserving AI agents for Status Communities. Local LLM inference, Logos Messaging transport, zero-leak architecture — no user data leaves the community's control.
v0.1.0 is a proof-of-concept that uses a cloud LLM (Anthropic API). Local inference is planned for v0.4.0.
Status Communities lack tools for knowledge management, onboarding, and coordination. Members ask the same questions repeatedly, decisions get lost in chat history, and new members have no guide.
| Approach | Problem |
|---|---|
| Discord/Telegram bots | All messages routed through a third party. One subpoena and the community's private discussions are exposed |
| "Just use ChatGPT" | Context is lost. No memory, no automation, no integration with the chat flow |
| Self-hosted AI | Requires DevOps + GPU. Messages must leave E2E encrypted channel to reach the server |
| Our approach | Agent lives inside the Status protocol — Logos Messaging transport, community-owned inference, existing E2E encryption |
https://github.com/epavlenko/status-app-agent/raw/main/docs/demo.mp4
uv sync # install dependencies
cp .env.example .env # configure (API keys, community ID)
./scripts/setup_status_backend.sh # start status-backend
uv run python -m src.bot # run the botStatus Community <-> status-backend (Go, HTTP/WS) <-> Python bot <-> Anthropic API (temporary, to be changed to local inference)
See docs/architecture.md for details.
| Version | Phase | Description |
|---|---|---|
| v0.1.0 | Phase 0 | Bot in Status Community, cloud LLM, chat context ✅ |
| v0.2.0 | Phase 1 | Token-based context, summarization, multi-user awareness |
| v0.3.0 | Phase 2 | RAG knowledge base from community discussions |
| v0.4.0 | Phase 3 | Local inference (Ollama), zero-leak architecture |
| v0.5.0 | Phase 4 | Multi-agent (Knowledge + Coordinator + Onboarding) |
| v0.6.0 | Phase 5 | Immutable Notes integration |
| v1.0.0 | Phase 6 | Full Logos Stack (Nomos, native Logos Messaging, payments) |
See docs/roadmap.md for the full vision and phased plan.
MIT