The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Most people don't need any config. ClawMetry auto-detects your workspace, logs, sessions, and crons.
If you do need to customize:
clawmetry --port 9000 # Custom port (default: 8900)
clawmetry --host 127.0.0.1 # Bind to localhost only
clawmetry --workspace ~/mybot # Custom workspace path
clawmetry --name "Alice" # Your name in Flow visualization
All options: clawmetry --help
Supported Channels
ClawMetry shows live activity for every OpenClaw channel you have configured. Only channels that are actually set up in your openclaw.json appear in the Flow diagram β unconfigured ones are automatically hidden.
Click any channel node in the Flow to see a live chat bubble view with incoming/outgoing message counts.
Auto-detection: ClawMetry reads your ~/.openclaw/openclaw.json and only renders the channels you've actually configured. No manual setup required.
Docker Deployment
Want to run ClawMetry in a container? No problem! π³
Quick start with Docker:
# Build the image
docker build -t clawmetry .
# Run with default settings
docker run -p 8900:8900 clawmetry
# Or with your OpenClaw workspace mounted
docker run -p 8900:8900 \
-v ~/.openclaw:/root/.openclaw \
-v /tmp/moltbot:/tmp/moltbot \
clawmetry
Note: When running in Docker, make sure to mount your OpenClaw workspace and log directories so ClawMetry can auto-detect your setup.
Requirements
Python 3.8+
Flask (installed automatically via pip)
OpenClaw running on the same machine (or mounted volumes for Docker)
Linux or macOS
NemoClaw / OpenShell Support
ClawMetry automatically detects NemoClaw β NVIDIA's enterprise security wrapper for OpenClaw that runs agents inside sandboxed OpenShell containers.
No extra configuration is needed in most cases. The sync daemon auto-discovers session files whether they live in ~/.openclaw/ on the host or inside an OpenShell container.
How it works
ClawMetry detects NemoClaw in two ways:
Binary detection β checks for the nemoclaw CLI and runs nemoclaw status to get sandbox info
Container detection β scans running Docker containers for openshell, nemoclaw, or ghcr.io/nvidia/ images, then reads sessions via volume mounts or docker cp
Session files synced from NemoClaw containers are tagged with runtime=nemoclaw and container_id metadata in the cloud dashboard, so you can tell them apart from standard OpenClaw sessions at a glance.
Recommended setup: sync daemon on the HOST
For the best experience, run ClawMetry's sync daemon on the host machine (not inside the sandbox). This avoids NemoClaw network policy restrictions.
# On the host (outside the sandbox)
pip install clawmetry
clawmetry connect
clawmetry sync
The sync daemon will automatically find sessions inside any running OpenShell containers.
Optional: explicit sandbox name
If auto-detection doesn't work, point ClawMetry at the right sandbox:
If you must run the sync daemon inside the OpenShell sandbox, add this egress rule to your NemoClaw network policy so it can reach the ClawMetry ingest API:
A Claude Code plugin that automatically captures everything Claude does during your coding sessions, compresses it with AI (using Claude's agent-sdk), and injects relevant context back into future sessions.
