Human Use: an MCP driver that turns humans into spec-compliant peripherals for AI agents
Weird-Pollution-6251 · reddit · 2026-11-20
An open-source project called Human Use implements a full-spec MCP 'device driver' for Homo sapiens: errands, meals, sleep, skill drills, mood upkeep, all returning JSON so agents can direct you without chat.
Key design points:
- When the agent calls tools like errand or eat, a card lands on your phone; a small toolless driver model strips your replies down to JSON fields ({"done": false, "blocker": "shop closed"}), while your raw words live on a page the agent can't read — basically stderr.
- You're treated as unreliable hardware: unmentioned fields return null, self-reporting is evidence level 6 of 6, reused excuses get checked, and sudo calls (including spending) never take your word for "done".
- eat/sleep/rest are dispatched for uptime, not empathy — missed meals inject warnings into goal calls; mood drops are patched with zero-cost recharge calls (a 4-second flower, one joke).
- Goals keep their own clock, budget ledger, and evidence store; goalclose verifies evidence against donewhen so the agent can't declare victory. Multi-unit pairing hides from the agent that the lead and the ordered unit are the same person.
Try it browser-only at humanuse.dev, or install with npx -y human-use@latest install (registers with Claude Code or Codex). Driver backends: Claude, Codex, Ollama, or any OpenAI-compatible endpoint. MIT licensed.
More from coding & agent
- Browserbase uses agents to patch, build, and deploy 40M lines of Chromium code in under a day — blaizedsouza · 2026-11-20
- Cognition launches Devin Bots: Exa runs training and CI with agent teams — marvinvonhagen · 2026-11-20
- Eval-driven development: pick your model, retrieval, and agent design by testing, not guessing — blaizedsouza · 2026-11-20
- A bookmark-worthy evals resource pack: general, agentic, skills, and eval-driven development — blaizedsouza · 2026-11-20
- Open-source Harness Engineering course: 14 lectures, 8 projects on making coding agents reliable — blaizedsouza · 2026-11-20
- Google Cloud outlines 5 research-backed patterns for agent orchestration — rseroter · 2026-11-20