Kobold MCP Server

Instructions

Implementation Reference

• src/index.ts:346-357 (handler)

  Generic handler for all POST-based tools, including kobold_complete. Validates input using the tool's schema, forwards the request to the KoboldAI API endpoint via makeRequest, and returns the JSON response.

  if (postEndpoints[name]) {
      const { endpoint, schema } = postEndpoints[name];
      const parsed = schema.safeParse(args);
      if (!parsed.success) {
          throw new Error(`Invalid arguments: ${parsed.error}`);
      }
  
      const result = await makeRequest(`${apiUrl}${endpoint}`, 'POST', requestData);
      return {
          content: [{ type: "text", text: JSON.stringify(result, null, 2) }],
          isError: false,
      };

• src/index.ts:113-119 (schema)

  Zod schema defining the input parameters for the kobold_complete tool, including prompt, max_tokens, temperature, top_p, and stop sequences.

  const CompletionSchema = BaseConfigSchema.extend({
      prompt: z.string(),
      max_tokens: z.number().optional(),
      temperature: z.number().optional(),
      top_p: z.number().optional(),
      stop: z.array(z.string()).optional(),
  });

• src/index.ts:266-268 (registration)

  Registers the kobold_complete tool in the ListTools response, providing its name, description, and input schema.

  name: "kobold_complete",
  description: "Text completion (OpenAI-compatible)",
  inputSchema: zodToJsonSchema(CompletionSchema),

• src/index.ts:343-343 (helper)

  Maps the kobold_complete tool to its KoboldAI API endpoint '/v1/completions' and references the input schema for validation.

  kobold_complete: { endpoint: '/v1/completions', schema: CompletionSchema },

• src/index.ts:146-162 (helper)

  Utility function that performs HTTP requests to the KoboldAI API, used by the tool handler to proxy requests and handle responses.

  async function makeRequest(url: string, method = 'GET', body: Record<string, unknown> | null = null) {
      const options: RequestInit = {
          method,
          headers: body ? { 'Content-Type': 'application/json' } : undefined,
      };
      
      if (body && method !== 'GET') {
          options.body = JSON.stringify(body);
      }
  
      const response = await fetch(url, options);
      if (!response.ok) {
          throw new Error(`KoboldAI API error: ${response.statusText}`);
      }
      
      return response.json();
  }