ClawMax, developed by Maximilien-ai, functions as a web orchestration layer for OpenClaw AI agents. The platform facilitates the management, monitoring, and orchestration of OpenClaw agent teams, workflows, and templates through a web-based interface. Key features include the ability to deploy agent teams, visualize workflow Directed Acyclic Graphs (DAGs), and track progress across an ecosystem.
The project emphasizes robust support for local and self-hosted models. It integrates OpenAI-compatible APIs, allowing users to configure and validate connectivity with services like LM Studio and other OpenAI-style endpoints that expose `/v1/models` and `/v1/chat/completions`. This enables the use of local models as distinct providers within the ClawMax environment, separate from official OpenAI models. Ollama is also supported in local and on-premise runtimes.
Recent updates have focused on improving image publication hardening, ensuring container images are smoke-pulled and run from GHCR before workflow success. Runtime diagnostics have been strengthened, with dashboard containers logging packaged dashboard versions and failing fast on version mismatches. Setup and onboarding processes have been streamlined, with `setup.sh` no longer requiring API keys or GitHub OAuth credentials during initial setup, and OpenClaw being installed automatically if missing.