loading…
Search for a command to run...
loading…
Enables AI assistants to control Anki Vector robots locally via natural language, providing tools for speech, motion, perception, and interaction without cloud
Enables AI assistants to control Anki Vector robots locally via natural language, providing tools for speech, motion, perception, and interaction without cloud dependency.
CI Python 3.11+ PyPI License: MIT Status: v1.0.0
Give your AI assistant a body.
VectorClaw is an MCP server that exposes Anki Vector robot capabilities as tools for AI assistants like OpenClaw — bridging natural language to real-world robot actions over a fully local, cloud-free stack.
┌─────────────┐ stdio MCP ┌──────────────────┐ gRPC/WiFi ┌─────────┐
│ AI Agent │ ─────────────── │ vectorclaw-mcp │ ─────────────── │ Vector │
│ (OpenClaw) │ │ (Python 3.11+) │ │ Robot │
└─────────────┘ └──────────────────┘ └─────────┘
All communication is local-only — no cloud dependency at runtime (Wire-Pod setup path).
See Security Architecture for the full trust model.
v1.0.0 Released · 2026-03-01
16 tools verified ✅ on hardware | 7 tools experimental ⚠️
Known limitations:
vector_drive_on_charger — activates cube but no reliable charger approachSee ROADMAP.md for the full milestone plan.
Requirements: Python 3.11+ · Wire-Pod running · Vector on local WiFi
For the complete walkthrough (Wire-Pod install, robot auth, WiFi config, troubleshooting) see docs/SETUP.md.
The vectorclaw-setup wizard handles configuration, SDK validation, connectivity
check, and a smoke test in one go:
pip install vectorclaw-mcp
vectorclaw-setup
You will be prompted for your robot's serial number and optional IP address. On success you'll see a clear SETUP PASSED message and the next-steps command. On failure every step includes an exact remediation hint.
See docs/OPENCLAW_SETUP_SKILL.md for full details.
Step 1 — Install VectorClaw
pip install vectorclaw-mcp
Step 2 — Configure Vector SDK
Wire-Pod is the canonical self-hosted server for Vector. Install the SDK distribution and run the one-time auth wizard:
pip install wirepod_vector_sdk
python -m anki_vector.configure
Note:
wirepod_vector_sdkinstalls under theanki_vectorPython namespace, so all imports and CLI commands useanki_vector.
The standalone anki_vector package requires working DDL cloud servers and is brittle on modern Python runtimes.
pip install "vectorclaw-mcp[legacy]"
python -m anki_vector.configure
Prefer wirepod_vector_sdk for reliable, cloud-independent operation.
Step 3 — Set environment variables
export VECTOR_SERIAL="your-robot-serial" # required — printed on underside of robot
export VECTOR_HOST="192.168.x.x" # optional — auto-discovered if omitted
Step 4 — Run the server
vectorclaw-mcp
# or
python -m vectorclaw_mcp
Add the following block to your mcporter.json (or equivalent MCP client config).
With uvx (recommended — no prior install needed)
{
"mcpServers": {
"vectorclaw": {
"command": "uvx",
"args": ["vectorclaw-mcp"],
"env": {
"VECTOR_SERIAL": "your-serial-here"
}
}
}
}
With pip install (if installed locally)
{
"mcpServers": {
"vectorclaw": {
"command": "vectorclaw-mcp",
"env": {
"VECTOR_SERIAL": "your-serial-here"
}
}
}
}
| Tool | Category | Description | Status |
|---|---|---|---|
vector_say |
🎙️ Speech | Make the robot speak text aloud | ✅ |
vector_animate |
🎭 Expression | Play a named animation | ⚠️ |
vector_drive_off_charger |
🏎️ Motion | Drive the robot off its charger | ✅ |
vector_drive |
🏎️ Motion | Drive straight and/or turn in place | ✅ |
vector_drive_on_charger |
🏎️ Motion | Drive Vector back onto its charger | ⚠️ |
vector_emergency_stop |
🏎️ Motion | Stop all motion immediately | ✅ |
vector_head |
🦾 Actuation | Set head angle (−22° – 45°) | ✅ |
vector_lift |
🦾 Actuation | Set lift height (0.0 – 1.0) | ✅ |
vector_look |
👀 Perception | Capture image from front camera | ✅ |
vector_capture_image |
👀 Perception | One-shot image capture | ✅ |
vector_face |
🖼️ Display | Display custom image on face screen | ✅ |
vector_scan |
🔍 Perception | Head scan for environment | ✅ |
vector_find_faces |
🔍 Perception | Scan for faces | ⚠️ |
vector_list_visible_faces |
🔍 Perception | List currently visible faces | ⚠️ |
vector_face_detection |
🔍 Perception | Get face detection summary | ⚠️ |
vector_list_visible_objects |
🔍 Perception | List currently visible objects | ⚠️ |
vector_cube |
🎲 Interaction | Interact with cube (dock/pickup/drop/roll) | ⚠️ |
vector_vision_reset |
👀 Perception | Disable all vision modes | ✅ |
vector_pose |
📍 Sensing | Get current position and orientation | ✅ |
vector_status |
📊 Status | Get battery level and charging status | ✅ |
vector_charger_status |
📊 Status | Get charger connection state | ✅ |
vector_touch_status |
📊 Status | Get touch sensor state | ✅ |
vector_proximity_status |
📊 Status | Get proximity sensor reading | ✅ |
Status legend: ✅ Verified on hardware | ⚠️ Experimental (limited/reliable issues)
⚠️ Charger prerequisite:
vector_driverequires the robot to be off the charger. Callvector_drive_off_chargerfirst, or setVECTOR_AUTO_DRIVE_OFF_CHARGER=1for automatic undocking.
See docs/MCP_API_REFERENCE.md for full parameter details and response schemas.
dev, use <type>/<short-description> naming (e.g. fix/vector-face-payload, feat/vector-scan)tests/; all tests use the mocked SDK — no hardware requiredpytest tests/ -v locally before opening a PR| Document | Description |
|---|---|
| Setup Guide | Wire-Pod install, robot auth, WiFi, SDK config, troubleshooting |
| Troubleshooting | Common runtime failures, smoke baseline, and escalation path |
| Runtime Support | Supported Python versions and CI policy |
| Document | Description |
|---|---|
| API Reference | MCP tool signatures, parameters, response schemas |
| Wire-Pod SDK Reference | Full SDK capability catalog |
| Wire-Pod SDK → MCP Integration Priorities | Now/Later/Skip decision table for future tools |
| Document | Description |
|---|---|
| Hardware Test Playbook | Repeatable on-robot validation protocol and PR checklist |
| Hardware Smoke Log | Running record of real-world smoke tests |
| Tool Docking Prerequisites | Which tools require undocked state |
| Document | Description |
|---|---|
| Security Architecture | Threat model, credential handling, input validation, network posture |
MIT — see LICENSE.
Run in your terminal:
claude mcp add vectorclaw -- npx Security
Low riskAutomated heuristic from public metadata — not a security guarantee.