Quoroom (quoroom-ai/room), an open-source swarm of autonomous AI agents, has been released. Its defining feature is distributed, collective decision-making run locally through Queen (strategy), Workers (execution) and Quorum (voting); it is available on GitHub .
Open Source · Autonomous AI Swarm
Quoroom: a swarm of AI agents that votes before it acts
An "earning-focused swarm intelligence engine" that runs a self-governing group of agents locally — a Queen sets strategy, Workers execute, and a Quorum makes the final call by vote. MIT-licensed and actively developed.
4
roles: Queen · Workers · Quorum · Keeper
3700
local dashboard port (localhost)
219
commits — actively updated repo
384
dim semantic-memory embeddings
How a decision flows through the swarm
Overall behavior emerges from local interactions — like an ant colony — aiming for more stable judgments than a single agent.
Queen
formulates strategy
→
Workers
execute delegated tasks
→
Quorum
votes: the final call
→
Keeper
human oversight & budget
Voting thresholds — pick how strict consensus must be
What's inside
Free local model
Ollama qwen3-coder:30b · local-only fail-closed mode
Paid APIs
OpenAI · Claude · Gemini via MCP
Wallet
EVM · USDC/USDT · Base, Ethereum, Arbitrum
Governance
Self-modification with audit logs & revert
The promise
Consensus among multiple agents may raise the robustness of judgments where a single agent tends to hallucinate or drift on complex tasks. Firms are already bringing multi-agent collaboration in-house.
The caution
Explicitly experimental software: complexity, cost and difficulty of control are open challenges. Users are told to monitor closely — and remain responsible for all agent actions.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…