Overview
This section is for engineering and integration teams who need robot capabilities to be extensible, diagnosable, and verifiable. The Ticos platform provides a suite of advanced tools designed to bridge the gap between high-level AI reasoning and physical hardware execution.

Core Capabilities
1. Extend Capabilities
Register and manage MCP (Model Context Protocol) tools. We support multiple integration patterns:
- SSE (Server-Sent Events): For cloud-to-cloud tool integration.
- STDIO: For tools running locally on the robot’s onboard computer.
- Streamable HTTP: For high-performance, long-running tool interactions.
2. Validate Integrations
Use the Lab environment to verify that your tool schemas match the robot’s perception capabilities.
- Schema Validation: Ensure JSON schemas for tool parameters are correctly interpreted by the LLM.
- Payload Inspection: View raw request and response payloads to identify data type mismatches or missing fields.
3. Debug & Diagnostics
Real-time tracing of the communication between the Ticos Cloud, the Agent, and physical drivers.
- Zenoh Tracing: Monitor the underlying message bus for latency or connectivity issues.
- Stream Diagnostics: Identify why video or sensor streams might be failing to reach Teleops.
Recommended Workflow
- Register: Add your custom tools in Tool Integration.
- Lab Validation: Verify the message flow and tool-call accuracy in the Real-time Debugger.
- Hardware Check: Run pre-release validation on physical hardware using Robot MCP.
- AI Alignment: Connect your preferred coding assistants via Working with AI Tools (MCP) to automate build and deployment steps.
Related Docs
- CLI — Command-line interface for platform management
- Pickle Parser — Efficiently parse Python Pickle robot data
- Authentication (Auth) — Unified auth based on Supabase
- Agentic Build — Automating robot software deployment
- Map Viewer — 2D/3D environment visualization and debugging