Index / TinyAGI · updated Oct 2, 2026
TinyAGI
TinyAGI/tinyclaw · healthy · rank 17 of 43 by stars
A multi-agent, multi-team orchestrator (formerly TinyClaw) aimed at running a 'one person company' of AI agents with isolated workspaces. Ships with a polished web portal (TinyOffice), team chat rooms, and Linear-style task management, but is still flagged experimental.
Facts
Repository
Runtime
Posture
Community
Security breakdown
Composite 62 / 100 · how these are scored
higher is safer
higher is safer
higher is safer
higher is safer
higher is riskier
Model access
Read from the repository, not written by a model · 32 files examined
1 provider · custom endpoint
Pinned models
gpt-5.3-codex released 2026-02-24 — 7 months old (from the public model catalogue)
Pin last edited 7 months ago
Evidence
Decision
Why choose TinyAGI over OpenClaw?
Why choose this
- Native multi-team orchestration with agent hand-offs and fan-out
- Built-in web portal (TinyOffice) with task/project management
- Multi-provider support (Claude, Codex, custom endpoints) with token management
Tradeoffs
- Experimental stability flag versus OpenClaw's more mature core
- Heavier Node.js runtime and daemon footprint
- Smaller ecosystem and fewer community skills/integrations
Best fit
- Solo founders wanting a team of collaborating AI agents
- Users who want a polished web dashboard for agent management
- Multi-channel assistant setups (Discord, WhatsApp, Telegram)
Avoid if
- You need a stable, production-hardened system (marked experimental)
- You want a lightweight single-agent assistant
- You avoid Node.js daemon/Docker operational overhead
Evidence is solid from README, commits, and release data showing active development, but community sentiment data is thin (1 Reddit match) and mostly indirect ecosystem mentions.
AI layer reviewed Sep 21, 2026 · how this is written
Star activity
3,620 stars today
Overview
TinyAGI (formerly TinyClaw) is a TypeScript-based multi-agent orchestrator positioned as an 'agent teams' platform for the one-person company. It runs multiple isolated AI agents with specialized roles that collaborate via chain execution and fan-out, coordinated through a SQLite-backed message queue with atomic transactions, retry logic, and dead-letter management. It supports Anthropic Claude, OpenAI Codex, and any OpenAI/Anthropic-compatible endpoint, with per-provider API key management built in.
Its standout differentiator is TinyOffice, a browser-based control plane featuring a pixel-office visualization of agents, persistent team chat rooms, Linear-style task and project management with comments, and unified services/providers/logs tabs. Agents connect to Discord, WhatsApp, and Telegram, and a plugin system allows custom message hooks and event listeners. Installation is a one-line curl script that auto-configures and launches the daemon.
Compared to OpenClaw, TinyAGI leans harder into multi-team collaboration and dashboard UX rather than single-assistant depth. However, it self-labels as experimental, and recent commits show rapid UI churn plus some security trade-offs (e.g., Docker running as root, relying on gVisor for isolation). It's best suited for early adopters who want an agent 'company' metaphor with a rich management UI.