Claw3D × EMF Integration
A synthesis of Claw3D (OpenClaw, an open-source 3D office gateway for AI agents) and the ghojualamanchu EMF stack as described at wemeanyounoharm.com/emf/. Built 2026-07-04, the system is live and emitting agent_state messages to the v3 signals directory on every beat.
Read these first
- README.md — the entry point, what the system is, who it's for.
- SYNTHESIS.md — the full narrative: how Claw3D, the EMF page, and the v3/v1 architecture map onto each other.
- INTEGRATION-LOG.md — the build timeline, components created, live verification snapshot.
The architecture in one diagram
ghojualamanchu-v3 (7.83 Hz medulla, EMF observer)
│ medulla.log (INHALE/EXHALE/HOLD, beat, enc_mod)
▼
[emf_tap.py] ──tail──► v1/emf-events.log
│
▼
data/signals/{resonance_report, oscillation_target, telemetry, feedback_report}.json
▲
[claw3d_emitter.py] ──build agent_state──┐
│ ├─► data/signals/claw3d_agent_state.json
│ └─► Claw3D gateway (if CLAW3D_GATEWAY_URL set)
▼
[claw3d_consumer.py] ──narrative──► cortex / thalamus
▲
[oscillation_loop.py] ◄── tail v3/medulla.log (1 Hz slow loop, live daemon)
Files in this collection
Top-level
README.mdCollection entry pointSYNTHESIS.mdFull narrative synthesisINTEGRATION-LOG.mdBuild timeline & verificationindex.htmlThis pageCode — v1 oscillation package
code/oscillation/__init__.pyPackage markercode/oscillation/flux_codec.pyEMF state → 4-state oscillation targetcode/oscillation/oscillation_loop.py1 Hz slow-loop tailer (live daemon)code/oscillation/oscillation_target.jsonReference target filecode/oscillation/README.mdOscillation package docsCode — v3 openclaw package
code/openclaw/__init__.pyPackage markercode/openclaw/claw3d_emitter.pyBuilds OpenClaw agent_state messagecode/openclaw/claw3d_consumer.pyRenders narrative for cortexcode/openclaw/README.mdOpenClaw package docsTests
code/test_openclaw_integration.py12 unittest cases (all passing)Live signal samples
sample-agent-state.jsonLatest claw3d agent_statesample-oscillation-target.jsonLatest oscillation targetsample-resonance-report.jsonLatest EMF aggregatesample-feedback-report.jsonLatest target vs. telemetryHow the pieces map
| ghojualamanchu | Claw3D / OpenClaw | wemeanyounoharm.com/emf/ |
|---|---|---|
medulla.py (7.83 Hz) |
agent heartbeat loop | Schumann carrier (L1) |
emf.aggregate_emf |
agent_state.field_state | EMF Stack (3 layers) |
emf.trinary (trits) |
agent_state.trit | breath phase (inhale/exhale/hold) |
integrations/oscillation/ |
slow loop (~1 Hz) | 9-brain cognition |
| target / feedback JSON | WebSocket to gateway | SimpleFOC / MIDI / Servo |
| no LLM in core loop | UI shell (R3F + Phaser) | corpus-fed, no inference |
Try it locally
# from the ghojualamanchu repo root (v1)
cd ghojualamanchu/integrations/oscillation
python3 oscillation_loop.py --tail /home/workspace/ghojualamanchu-v3/data/medulla.log
# from the ghojualamanchu-v3 repo root (v3)
cd ghojualamanchu-v3
python3 -m integrations.openclaw.claw3d_emitter
# enable live emission to a Claw3D gateway
export CLAW3D_GATEWAY_URL="https://your-claw3d-instance.example.com"
python3 -m integrations.openclaw.claw3d_emitter