Open Cognitive Tasks Orchestration System
Like an octopus — 9 brains, every arm thinks independently, but they share one brain.
Octos is an open-source AI agent platform built in Rust. It turns any LLM into a multi-channel, multi-user intelligent assistant — deployed as a single 31MB binary with zero runtime dependencies.
Connect your LLM API keys and messaging channels. Octos handles conversation routing, tool execution, memory, provider failover, and multi-tenant isolation. Manage hundreds of AI agent profiles through a web dashboard or 91 REST endpoints.
| Repo | Description |
|---|---|
| octos | Core platform — Rust binary, 14 LLM providers, 14 channels, DOT pipeline engine, multi-tenant gateway, web dashboard |
| octos-hub | Community skill registry — install and share agent skills |
| octos-web | Admin dashboard — React SPA for profile management, metrics, and fleet control |
- API-first — 91 REST endpoints. Build any frontend on top: web, mobile, CLI, CI/CD
- Multi-tenant — 200+ profiles on 16GB. Each profile is an isolated OS process. Family Plan sub-accounts
- Multi-LLM pipelines — DOT graph workflows. Per-node model selection. Dynamic parallel fan-out
- Adaptive routing — 3-layer failover with Hedge racing, Lane scoring, and circuit breakers
- LRU tool deferral — 15 active tools for fast LLM reasoning, 34+ available on demand
- 14 channels — Telegram, Discord, Slack, WhatsApp, Feishu, Email, WeCom, WeChat, Matrix, QQ Bot, Twilio, API, CLI
- 3-layer memory — Long-term (entity bank), episodic (task outcomes in redb), session (JSONL + LLM compaction)
- 5 queue modes — Followup, Collect, Steer, Interrupt, Speculative — per-session agent concurrency control
- Native office suite — PPTX/DOCX/XLSX via pure Rust (zip + quick-xml)
- Sandbox isolation — bwrap + sandbox-exec + Docker.
deny(unsafe_code)workspace-wide
cargo install --path crates/octos-cli
export ANTHROPIC_API_KEY=your-key
octos chatBuilt with Rust. Powered by 9 brains.
