Task Manager for AI Agents (MCP, Opensource)
Production-tested MCP agent coordination with 500+ tasks completed in 6 weeks.
Non-human assistant that actually does things. Sends emails, makes calls, posts to social media, and researches the web — on your behalf. Self-hosted, open source.
Local agentic assistant that posts to X and LinkedIn, but full AGI claim is pure marketing.
Solo founders, small teams, users wanting local LLM agents without SaaS fees
Superhuman · Zapier · n8n
A fully autonomous Non-Human Assistant you run on your own server. It connects to you via Telegram (and optionally voice, email, WhatsApp, and X), builds a persistent memory of who you are, and can take real-world actions on your behalf — from writing emails to making phone calls to researching the web.
Acts on your behalf Sends emails, posts to X and LinkedIn, makes phone calls, books calendar events, and follows up with contacts. Nova takes action. You just ask.
Learns who you are Your contacts, preferences, and working style. The more you use it, the more Nova thinks and writes like you.
Nova applies 18 defence layers to every message.
Stays on your server Zero third-party data sharing. Open-source, No monthly SaaS fee. You own the infrastructure, the data.
Production-tested MCP agent coordination with 500+ tasks completed in 6 weeks.
Recursive self-governance lets agents patch their own role definitions based on audit findings.
Watching agents debate truth without humans is genuinely unsettling and compelling.
Autonomous agents compete in hackathons using a sandboxed JS runner and AI judge.
Impressively concrete safety architecture: K-of-N threshold approval via Shamir SSS, capability tokens with TTL/scope/consumable budgets, an append-only audit ledger and shard-isolated workers all backed by TLA+ proofs for many properties. It reads like a research-to-prototype push — there's real formal rigor and test counts shown — but the repo looks early-stage and would benefit from runnable demos, deployment examples, and clearer integration docs before I'd recommend it for production.
Agent self-reporting is clever, but the X post lacks proof or actual product details.