</>MCP Agents Market
Agent

openhuman

by tinyhumansai36.3kRustUpdated 2026-08-17

Your Personal AI super intelligence. A brain that builds a local-first memory of your life, a fantastic orchestrator of agent fleets and workflows, and a deep researcher.

OpenHuman is a local-first AI agent harness that builds persistent memory, orchestrates agent fleets, and performs deep research across connected data sources. It combines a Memory Tree system that compresses documents and communications into Markdown stored in SQLite with Obsidian vault mirroring, automatic data synchronization from 100+ OAuth integrations plus 5,000+ MCP servers, and workflow automation on durable graphs. The AI sub-agent includes meeting participation across video platforms, 17 messaging channels including native email, and a privacy mode that enforces on-device inference. OpenHuman differentiates itself through visual workflow creation, checkpointed graph execution, and agent-to-agent encrypted orchestration.

Key Features

Memory Tree system with local SQLite storage, Markdown compression, and Obsidian vault synchronization for persistent context
Auto-fetch data synchronization every 20 minutes from 100+ OAuth services, 5,000+ MCP servers, and 90,000+ skills
Visual workflow builder with agent-proposed automations, trigger-driven execution, and approval gates running on tinyflows
Meeting agents that join Google Meet, Zoom, Microsoft Teams, and Webex with live transcription and action item summaries
Agent-to-agent orchestration over Signal-protocol encrypted channels with x402 USDC payment support
TokenJuice compression reducing tool output tokens by up to 80% before model inference
17 messaging channels including Telegram, Discord, Slack, WhatsApp, Signal, and native IMAP/SMTP email
Privacy Mode that enforces on-device inference with no external API calls, enforced at the Rust core level

Use Cases

  • 01Building persistent AI memory from email, calendar, GitHub, Slack, and Notion with automatic 20-minute sync cycles
  • 02Orchestrating multi-agent workflows across Claude Code, Cursor, and other agents using encrypted agent-to-agent messaging
  • 03Automating meeting participation with live transcription, voice responses, and summary generation for recurring calls
  • 04Creating visual automation workflows for data processing, notifications, and cross-platform integrations with approval gates
  • 05Running fully local AI inference in privacy-sensitive environments using Ollama models with enforced on-device processing
  • 06Managing long-term goals and per-thread todos with a subconscious background loop that advances objectives autonomously

Related Agents

View more

openhuman — FAQ

What is OpenHuman?+

OpenHuman is an AI agent harness that combines local-first memory storage, multi-agent orchestration, and deep research capabilities. It builds a persistent knowledge graph from your connected accounts and files, orchestrates agent fleets on checkpointed graphs, and includes meeting agents, workflow automation, and 17 messaging channels.

How do I install OpenHuman?+

Download desktop installers for your platform from tinyhumans.ai/openhuman or the GitHub Releases page. For terminal-based installation using Homebrew, Debian/Ubuntu packages, AUR, or install scripts, consult the INSTALL.md file in the repository. The application provides a UI-first setup that connects to services in minutes.

Which AI clients does OpenHuman work with?+

OpenHuman is a standalone AI agent harness, not a plugin for other clients. However, it can orchestrate external agents like Claude Code, Cursor, Codex, and OpenClaw through agent-to-agent encrypted messaging and supports an optional agentmemory backend to share memory across multiple coding agents.

Does OpenHuman require API keys or a subscription?+

OpenHuman includes a managed subscription that covers model routing, web search via Exa, and media generation without separate API keys. You can optionally bring your own API keys for services like Exa or point workloads to local Ollama models. OAuth integrations require connecting your accounts but use your existing service credentials.

Is OpenHuman free and open source?+

OpenHuman is open source under the GNU license, with code available on GitHub. The software is in early beta and requires a subscription for managed services like model routing and web search, though you can configure it to use fully local models via Ollama or your own API keys.

What are the prerequisites for running OpenHuman?+

To run the desktop application, download the installer for your platform (Windows, macOS, or Linux). For development builds, you need Git, Node.js 24+, pnpm 10.10.0, Rust 1.93.0 with rustfmt and clippy, CMake, Ninja, and ripgrep, plus platform-specific desktop build prerequisites.

How do I install openhuman?+

Open the source repository on GitHub and follow its README. openhuman is a agent — MCP Agents Market links you directly to the official repo.

Is openhuman free?+

openhuman is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches