What is markus?

Markus is an open-source AI workforce platform that runs entirely on your own hardware. Instead of a single chatbot, it creates teams of specialized AI agents that plan, execute, and peer-review work autonomously — from research and content operations to software development and monitoring. You provide a high-level goal in natural language; Markus assembles the team, breaks down tasks, runs them in parallel with persistent memory, and gates every delivery through peer review. It runs as a desktop app (macOS, Windows, Linux) or via CLI on a server, with a web UI at localhost:8056. All data stays in a local SQLite database; you bring your own LLM API keys (Anthropic, OpenAI, Google, DeepSeek, Ollama, OpenRouter, etc.) or use local models for fully offline operation. The project is pre-1.0 (0.11.x) with frequent releases and breaking changes possible. A free tier with trial credits is available; paid plans add managed multi-model access, higher quotas, remote access, and multi-user collaboration.

Should you self-host markus?

markus scores 48/100 on selfhost.directory (how scoring works) and ranks #110 of 209 in AI Agents. It has 209 GitHub stars.

It shipped 6 stable releases in the past year. The latest, v0.12.0, came out on 27 October 2026. New versions land roughly every 2 weeks.

Setup is moderate: expect some configuration beyond a single command. It is Apache-2.0-licensed and written mostly in Typescript.

Its closest alternative on selfhost.directory is Orca (91/100, 108 642 stars). It scores higher overall, so compare the two before you pick. See markus vs Orca side by side.

Features

Agent-to-agent protocol
Autonomous Agents
Bring Your Own Key (BYOK)
Communication bridges
Fully offline mode
GUI and browser automation
Local-first data
Multi-agent Teams
Multi-Provider LLM Support
Peer review workflow
Persistent memory
REST + WebSocket API
Skill ecosystem
Web UI dashboard

FAQ

How do I upgrade?

Desktop: download the latest installer from GitHub Releases and reinstall. npm: npm install -g @markus-global/cli. From source: git pull && pnpm install && pnpm build.

How do I configure LLM models?

Everything is done visually in the Web UI Settings page — no config files or env vars. Any common provider works: Anthropic, OpenAI, Google, DeepSeek, MiniMax, Ollama, SiliconFlow, Moonshot, etc. Just add the provider and API key.

Where is the data stored? Do I need a separate database?

No. The platform uses a local SQLite database out of the box. The Hub marketplace stores its DB at ~/.markus-hub/hub.db by default (auto-created and migrated on first start).

What if a port is already in use?

Backend API defaults to port 8056; Web UI is bundled. From source: backend API :8056, frontend Vite Dev :8057. Ports can be adjusted in the Settings page of the Web UI — no config file edits required.

What are the default login credentials, and how do I change them?

After the first launch, sign in with admin@markus.local / markus123. The onboarding wizard will prompt you to set your real name, email, and a new password.

How do I get support?

Report bugs and request features on GitHub Issues. For real-time help and discussion, join the community on Feishu.

Can I self-host Markus?

Yes. Markus Core is open source under Apache-2.0. You can self-host it on any machine — macOS, Linux, or Windows. Enterprise plans offer private deployment support and custom SLAs.

Is my data private?

Markus runs entirely on your local machine. Your projects, agent memory, and conversations stay on your device. No data is sent to Markus servers unless you explicitly use optional cloud features like Markus Hub.

Which AI models are supported?

Markus supports Anthropic (Claude), OpenAI (GPT), Google (Gemini), DeepSeek, MiniMax, Ollama (local models), OpenRouter, Qwen, Seedance, and any OpenAI-compatible API. Switch providers in one line of config — no lock-in.

What's the difference between Free and paid plans?

All plans include unlimited agents, teams, and projects. The free tier (with your own API keys) will always be available. Paid plans unlock managed multi-model access, optimized model routing, higher credit quotas, premium community assets, the ability to publish and earn, remote access, and multi-user collaboration.

What does "Bring Your Own Key" mean?

You provide your own API keys from LLM providers like OpenAI, Anthropic, Google, or DeepSeek. Markus uses them to power your AI agents. You only pay the provider directly for the tokens you consume — Markus charges nothing on top.

Is Markus free?

Yes. The free tier includes trial credits or you can use your own API keys (Bring Your Own Key). Unlimited agents, teams, and projects are included. A free tier will always be available.