Frontrun MCP Server

Instructions

Implementation Reference

• index.js:368-374 (handler)

  The handler function for frontrun_classify tool that takes usernames and/or twitter_user_ids, constructs a request body, and calls the POST /classify API endpoint, returning the result as JSON text.

  async ({ usernames, twitter_user_ids }) => {
    const body = {};
    if (usernames) body.usernames = usernames;
    if (twitter_user_ids) body.twitter_user_ids = twitter_user_ids;
    const result = await apiCall('POST', '/classify', body);
    return { content: [{ type: 'text', text: JSON.stringify(result, null, 2) }] };
  }

• index.js:364-367 (schema)

  Zod schema defining the input parameters for frontrun_classify: optional arrays of usernames and twitter_user_ids, both as strings.

  {
    usernames: z.array(z.string()).optional().describe('Usernames to classify'),
    twitter_user_ids: z.array(z.string()).optional().describe('Twitter user IDs to classify'),
  },

• index.js:361-375 (registration)

  Registration of the frontrun_classify tool with the MCP server using server.tool(), including the tool name, description, schema, and handler.

  server.tool(
    'frontrun_classify',
    'Run classification on specific entities. Returns AI classification merged with your custom rules and tags. Use this to analyze entities on demand.',
    {
      usernames: z.array(z.string()).optional().describe('Usernames to classify'),
      twitter_user_ids: z.array(z.string()).optional().describe('Twitter user IDs to classify'),
    },
    async ({ usernames, twitter_user_ids }) => {
      const body = {};
      if (usernames) body.usernames = usernames;
      if (twitter_user_ids) body.twitter_user_ids = twitter_user_ids;
      const result = await apiCall('POST', '/classify', body);
      return { content: [{ type: 'text', text: JSON.stringify(result, null, 2) }] };
    }
  );

• index.js:29-73 (helper)

  Helper function used by frontrun_classify (and other tools) to make authenticated API calls to the Frontrun API with error handling, timeout, and response parsing.

  async function apiCall(method, path, body = null) {
    const url = `${API_URL}/v1${path}`;
    const options = {
      method,
      headers: {
        'X-API-Key': API_KEY,
        'Content-Type': 'application/json',
      },
    };
    if (body) {
      options.body = JSON.stringify(body);
    }
  
    const controller = new AbortController();
    const timeout = setTimeout(() => controller.abort(), 60000);
    options.signal = controller.signal;
  
    let response;
    try {
      response = await fetch(url, options);
    } catch (err) {
      clearTimeout(timeout);
      if (err.name === 'AbortError') return { error: 'Request timed out (60s). Try a narrower query.' };
      return { error: `Network error: ${err.message}` };
    }
    clearTimeout(timeout);
  
    if (response.status === 429) {
      const retry = response.headers.get('Retry-After') || '60';
      return { error: `Rate limited. Retry in ${retry}s.` };
    }
    if (response.status === 401) {
      return { error: 'Invalid API key. Check FRONTRUN_API_KEY.' };
    }
    if (response.status === 402) {
      const data = await response.json();
      return { error: 'Insufficient balance', ...data };
    }
    if (!response.ok) {
      const text = await response.text();
      return { error: `HTTP ${response.status}: ${text.slice(0, 500)}` };
    }
  
    return response.json();
  }