Open Multi-Agent

Open Multi-Agent

Self-hosted agent runtime with approvals and audit

Description

Letting AI agents act inside a company is scary when they might do something they shouldn't and leave no clear trail. OMA is a self-hosted TypeScript agent runtime where consequential actions wait for durable approvals and every run leaves a record verifiable offline, byte for byte.

No telemetry and no hosted control plane; use cloud, Ollama, vLLM or Chinese providers, and scaffold PR review or security agents in one command.

Features



Approvals:Humans confirm key actions.

Verifiable runs:Tamper-evident records.

Any model:Cloud, local or Chinese providers.

Scaffolding:One-command agents.