Tencent Releases Octop, a Self-Hosted Multi-User AI Platform in One Process
Tencent Cloud open-sourced Octop, a single-process, self-hosted AI assistant that gives every user their own team of MBTI-flavored agents across web, CLI, and chat apps.
- Tencent Cloud open-sourced Octop, an MIT-licensed self-hosted multi-user, multi-agent AI assistant.
- Single Python process serves web dashboard, CLI, IM bridges, and cron from one SQLite database.
- Built on the Harness stack: LangGraph agent runtime, IM gateway, portable memory, CDP browser automation.
- 16 MBTI persona templates plus JWT multi-user isolation, tool approval, and shell guardrails.
- Bidirectional ACP: expose agents to Zed or OpenCode, delegate to Claude Code, Codex, CodeBuddy.
- Install via
uvone-liner, PyPI, or Docker Compose; supports OpenAI APIs, Qwen, Ollama.
Tencent Cloud’s Octop packs a multi-user AI platform into one process
Tencent Cloud has published Octop on GitHub, an MIT-licensed, self-hosted AI assistant for households and small teams. It combines a web dashboard, command-line interface, messaging bridges, browser automation, and scheduled jobs while giving each user separate agents, workspaces, credentials, and provider settings. Developers can deploy those features on one host without operating Redis, a message broker, or an external job queue.
One process, many entry points
Octop routes browser chats, CLI requests, messaging events, and scheduled tasks through an in-process HarnessProcessor. SQLite in write-ahead logging mode stores control-plane state in ~/.octop/octop.db, allowing the service to reconstruct its runtime state after a restart.
| Layer | Implementation |
|---|---|
| Interfaces | Web dashboard, CLI, Feishu, DingTalk, QQ, Discord, WeCom, and other messaging channels |
| Request pipeline | One in-process HarnessProcessor |
| Control plane | SQLite through aiosqlite, using WAL mode |
| Scheduling | APScheduler cron jobs |
| Deployment model | Single Python process on a host, service manager, or Docker Compose |
The compact architecture reduces deployment work and concentrates operational risk. A process crash or host outage affects every interface, scheduled jobs pause while the service is unavailable, and horizontal scaling is outside the current design.
Inside the Harness stack
Tencent describes Octop as a composition of four Harness components:
This story is for Pro members
You've reached the end of the free preview. Upgrade to AlphaSignal Pro to read the full article - and everything else behind the paywall.
Comments
