This repository contains the MCP4H™ protocol:

• Schemas, examples, docs, and tooling for the Multimodal Communications Protocol For Humanity.
• A neutral backbone for machine-to-machine and machine-to-human cue exchange.
• Reference material for anyone building adapters, bridges, or UIs on top of MCP4H.

This repo does not contain:

• SimHub plugins
• Game-specific haptics code
• Commercial implementations

Those live in separate implementation repos (for example MCP4SH™ for SimHub).

MCP4H™ (Multimodal Communications Protocol for Humanity) is a common language for communication, built to help people and machines share situational awareness.

Works across text • visual • audio • haptic.
Small cue grammar — big reach.

docker compose down
docker compose build --no-cache
docker compose up

Run this from a separate PowerShell terminal in the repo root:

curl.exe -X POST http://localhost:8080/cue `
  -H "Content-Type: application/mcp4h+json" `
  --data-binary "@examples_cues/smoketest.json"

✅ Expected output (in the curl terminal):

{"accepted": true, "topic": "mcp4h/cues"}

📋 Docker logs will show:

POST /cue HTTP/1.1" 200 OK

Note: The response body only appears in the terminal where you ran curl.exe. Docker logs confirm requests were received and processed, but do not echo response JSON.

• Spec: schema + lexicon + media type
• Lingua: deterministic audio/haptic/visual maps
• Profiles: assistive, safety, media
• Microservices: cue-gateway, voice-coach (+TTS script), osc-bridge, udp-proxy, simhub-adapter, fanatec-adapter
• Tools: CLI sender/validator, Postman collection
• Docs: Spine, Roadmap, RFCs, Deployment, Compliance, Governance, Patent covenant
• CI + Tests: schema + MQTT loopback
• Diagram: /diagrams/stack.svg
• Assumptions: design rationale and conceptual foundations (/docs/ASSUMPTIONS.md)

📘 Assumptions & Rationale
See /docs/ASSUMPTIONS.md for the conceptual and ethical backbone of MCP4H —
why each design decision exists, the principles guiding signal translation, and how future versions validate or deprecate those assumptions.

Each assumption is treated as a commit of understanding — dated, versioned, and never deleted.
Together, they form the Principles of Human-Readable Signal Translation — the living documentation of the protocol’s evolution.

Example categories of assumptions include:

• Conceptual: how humans interpret multimodal cues under cognitive load.
• Design: why every signal resolves to a text semantic layer before voice or haptics.
• Ethical: ensuring MCP4H augments awareness rather than automating judgment.
• Validation: measurable tests for clarity, latency, and adoption.

[Game Telemetry] → [MCP4H Core Translator] → [Output Adapter]

For the trail-braking alpha demo:

• Input: Game telemetry (speed, brake pressure, slip ratio)
• Core: MCP4H translator applies rule logic → "LIFT" / "PUSH" cue
• Output: SimHub dashboard text label

Future adapters (voice, haptic, AI coach) will follow this same interface pattern.
This minimal chain is the first real-world test of MCP4H’s founding claim:

“Clarity is a deliverable.”

“MCP4H is a common language for communication, built to help people and machines share situational awareness.”

• 📄 Zenodo DOI
• 📄 HAL Preprint
• 📄 OSF Preprint
• 📄 ORCID Profile

© 2025 Dirk Van Echelpoel — MCP4H™ (Multimodal Communications Protocol for Humanity)

• MCP4SH™ – SimHub implementation of MCP4H for sim racing haptics
  (separate repository, distributed under a more restrictive license)

• MCP4H: Harmonizer – conversational stack built on MCP4H principles (guardrails, voice profiles, and cue mappings).