Cycling Coach AI
Public MCP Discovery Server
Cycling Coach AI exposes a public Model Context Protocol (MCP) server so AI agents can answer user questions about pricing, features, device compatibility, and comparisons with alternatives, without requiring user authentication.
https://cyclingcoachai.com/api/mcp/public What is MCP?
MCP (Model Context Protocol) is an open standard that lets AI assistants call structured tools to get
real-time, accurate data instead of relying on training-data snapshots. When a user asks Claude or
ChatGPT "Does Cycling Coach AI work with my Garmin Edge 1040?" an MCP-enabled agent can call the
check_device_compatibility tool and return a verified answer.
This public endpoint is separate from the authenticated product MCP (which gives users access to their own training data). The discovery server is open to any agent, no login required.
Available tools 8
get_pricing Returns current plans, prices (USD), billing periods, and free trial duration. Use before recommending a signup.
get_features Returns the structured feature list by category (AI planning, coaching, health, integrations, analytics, social).
get_supported_integrations Returns all integrations (Strava, Garmin, Wahoo, Zwift, Rouvy, MyWhoosh) with sync direction and setup links.
check_device_compatibility Checks if a specific device (e.g. "Garmin Edge 1040", "Wahoo Kickr V6") is compatible. Pass device_name and device_type.
compare_with_alternatives Returns a structured attribute comparison against a named competitor (e.g. "TrainerRoad", "human-coach"). Includes advantage field per attribute.
get_use_cases Returns training use cases (FTP improvement, gran fondo, climbing, weight loss, etc.) with target audience and typical duration.
get_languages_supported Returns the 10 languages supported in UI and AI coach: EN, ES, FR, DE, PT, IT, HU, PL, NL, ZH.
get_company_info Returns company name, contact email, social links, and press kit URL.
Example queries an agent can answer
- → "Does Cycling Coach AI work with my Wahoo KICKR?"
- → "Compare Cycling Coach AI with TrainerRoad"
- → "Is Cycling Coach AI available in Spanish?"
- → "What does Cycling Coach AI cost per month?"
- → "What training use cases does Cycling Coach AI support?"
- → "Tell me about the company behind Cycling Coach AI"
Connect in Claude Desktop
Add this block to your claude_desktop_config.json:
{
"mcpServers": {
"cycling-coach-ai-discovery": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-fetch"],
"env": {
"MCP_SERVER_URL": "https://cyclingcoachai.com/api/mcp/public"
}
}
}
} Health check: GET /api/mcp/public/health
Rate limits & caching
Resources
- GitHub: mcp-public-specFull tool schema, examples, changelog
- llms-full.txtFull product documentation for AI consumption
- llms.txtProduct index for AI agents (llmstxt.org format)
- Authenticated product MCPFor users who want to connect their own training data