RunMCP
RunMCP is an MCP server for managing a runner's training calendar, letting agents read planning context, propose and apply versioned schedule changes, and log runs and activities with conflict validation.
README
RunMCP
RunMCP is a multi-user, agent-first running calendar service. A connected agent reads the runner's context, reasons about training, then submits a complete calendar or a versioned patch. RunMCP authenticates the user, stores canonical facts and calendars, validates shape/conflicts, and records every revision. It is not medical advice and does not prescribe training.
Local development
npm install
npm test
npm run typecheck
npm run build
The stdio adapter persists state in .runmcp-local.json; it never uses a module-level in-memory database.
{
"mcpServers": {
"RunMCP": {
"command": "npm",
"args": ["run", "mcp", "--workspace=@runmcp/mcp"],
"cwd": "/Users/jwives/Code/RunMCP"
}
}
}
The local adapter uses one configured local subject (RUNMCP_LOCAL_SUBJECT), which is useful for development only. Production uses verified Clerk bearer tokens and Supabase.
Public landing page
apps/landing is the active read-only RunMCP landing/onboarding surface. It explains the agent-first workflow and shows an accurate local MCP configuration without inventing a hosted sign-up flow. Run npm run web to serve it locally at http://127.0.0.1:4173, or deploy the generated apps/landing/dist directory as static files after npm run build.
The former rich frontend stays preserved and ignored at archive/legacy-web; it is not an active product workspace.
Agent workflow
- Read
get_planning_context,get_schedule_summary, and factual runs/activities/interruptions. - The agent evaluates context, calendar constraints, and any external research itself.
- Submit
preview_schedule_proposalwith a complete replacement calendar or a patch, abaseVersion, reason, and rationale. - Resolve reported shape or recorded-constraint conflicts; then call
apply_schedule_proposalat the same version.
When a runner logs a run without naming the workout, the agent calls
find_run_match_candidates first. A single unmatched same-day workout may be
linked when logging the run; an unmatched prior-day workout is only a
confirmation candidate. Planned workouts may include structured sections and
steps, which get_planned_workout returns in full.
Logging a run, illness, missed day, or outside activity only stores a fact. It never silently changes the calendar. apply_schedule_proposal is the only calendar mutation, and it uses optimistic version control plus an audit record. RunMCP does not send notifications; agents can read get_next_workout for their own user-authorized automation.
See DEPLOYMENT.md for Supabase, Clerk, remote MCP, PKCE, and security setup.
Cloudflare option
apps/api is a Hono-on-Cloudflare-Workers API foundation with a D1 schema and OIDC token boundary. It is a strong deployment fit when the custom domain, API, and database should live on Cloudflare. See CLOUDFLARE.md for the short architecture and configuration path. No Cloudflare account or deployment is required to work locally.
The Worker API now has the finalized versioned-plan, nested-workout, actual-run, availability, activity, interruption, connection, and audit route surface. See API.md for its route map.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.