omnigent

What it is and what it does

Omnigent is a meta-harness providing unified orchestration over multiple AI agents—Claude Code, Codex, Cursor, and custom YAML-defined agents—letting you swap or combine them without rewriting code. It runs on Python 3.12+ and depends on FastAPI, SQLAlchemy, openai, and OpenTelemetry for server, session management, and observability.

You can start sessions in terminal and continue in browser or phone with files in sync, supervise multiple agents in one session, enforce policies (pause for approval, cap spend, limit tools), run agents in cloud sandboxes, and collaborate by sharing sessions. The package includes CLI and web UI; macOS and Linux offer native terminal wrappers, while Windows runs in degraded mode without filesystem sandboxing or native PTY wrappers.

Use it for:

• Orchestrate multiple specialized agents in one session to divide work by capability.
• Run agent workloads in disposable cloud sandboxes without a local laptop.
• Enforce governance policies across all agents—pause before risky actions, cap spend, restrict tools.
• Collaborate with teammates by sharing a live agent session for co-driving or forking.
• Start an agent task in terminal, pick it up in browser or phone, keep messages synced.
• Define custom agents declaratively in YAML and mix them with first-party agents.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Omnigent is an open-source orchestration layer that lets you run, supervise, and collaborate on multiple AI agents (Claude, Codex, Cursor, custom YAML-defined agents) from terminal, browser, or phone, with policy enforcement and cloud sandbox support.

Yes, with conditions. Omnigent is actively maintained (released 3 days ago), has no known vulnerabilities, uses permissive Apache 2.0 license, and solves multi-agent orchestration across devices. Install friction is low (pure Python wheel), but requires Python 3.12+, uv, git, Node.js 22+, tmux, and bubblewrap on Linux. Windows gets reduced functionality. Worth installing if you need multi-agent orchestration or cross-device session sync.

Install

omnigent on PyPI

pip install omnigent

uv add omnigent

poetry add omnigent

Installing omnigent

Before you install

Low friction: pure Python wheel with no compiled dependencies. Active maintenance—released 3 days ago. Requires Python 3.12+ and several system tools (uv, git, Node.js 22+, tmux, bubblewrap on Linux) that the installer can provision; Windows support is degraded.

License in practice

Apache 2.0 permissive license: you can use, modify, and distribute freely in commercial and private projects, with no copyleft obligations.

Quickstart

# Install via installer (recommended)
curl -fsSL https://raw.githubusercontent.com/omnigent-ai/omnigent/main/scripts/install_oss.sh | sh

# Or manual install with uv
uv tool install omnigent

# Start the server and CLI
omnigent server &
omnigent run agent.yaml

Requires Python 3.12+, uv, git, Node.js 22 LTS or newer with npm and pnpm, tmux, and bubblewrap (Linux only). Windows has reduced functionality.

Verify before relying

• Whether all 41 runtime dependencies are required for basic usage or only for optional integrations.
• Performance characteristics when supervising multiple agents concurrently.
• Real-time collaboration latency and scalability limits.
• Exact policy enforcement capabilities and their interaction with different agent types.

Package facts

Evidence: omnigent-0.9.0-py3-none-any.whl

Keywords: agents, llm, ai, databricks, cli

Tags

More Artificial Intelligence packages

Further reading