TPMJS vs Manual Function Calling

The Per-Provider Schema Problem

Each LLM provider expects tool schemas in a different format. A single tool requires separate definitions for each provider:

{
  "type": "function",
  "function": {
    "name": "search_web",
    "description": "Search the web",
    "parameters": {
      "type": "object",
      "properties": {
        "query": {
          "type": "string",
          "description": "Search query"
        },
        "limit": {
          "type": "number",
          "description": "Max results"
        }
      },
      "required": ["query"]
    }
  }
}

{
  "name": "search_web",
  "description": "Search the web",
  "input_schema": {
    "type": "object",
    "properties": {
      "query": {
        "type": "string",
        "description": "Search query"
      },
      "limit": {
        "type": "number",
        "description": "Max results"
      }
    },
    "required": ["query"]
  }
}

Two providers, one tool — already two schema definitions to maintain. Add the handler function, validation logic, error handling, and tests, and each tool costs 1-4 hours. Multiply by ten tools and three providers: 30 schema definitions, 30 validation functions, and a maintenance surface that grows linearly.

How TPMJS Eliminates Schema Boilerplate

TPMJS tools define their schema once using Zod — the same Zod object used for runtime validation is also the schema definition. The TPMJS enrichment pipeline auto-extracts the JSON Schema from the Zod definition, so there's no separate schema to write or maintain:

// This is the entire tool definition.
// Zod schema = input validation = JSON Schema = MCP inputSchema

import { tool } from 'ai';
import { z } from 'zod';

export const searchWeb = tool({
  description: 'Search the web and return results',
  parameters: z.object({
    query: z.string().describe('Search query'),
    limit: z.number().optional().describe('Max results'),
  }),
  execute: async ({ query, limit = 10 }) => {
    // Implementation — no separate schema, no per-provider format
    const results = await fetch(`https://api.example.com/search?q=${query}&limit=${limit}`);
    return results.json();
  },
});

// What happens automatically:
// 1. npm publish → TPMJS changes feed picks up the package
// 2. Enrichment sync extracts inputSchema from Zod via the executor
// 3. Quality score calculated: tier + log(downloads) + log(stars) + metadata
// 4. Available via MCP, CLI, web search, and AI SDK

Feature Comparison

At Scale

Built-In Quality Signals

With manual function calling, there's no way to know if a tool implementation is reliable. TPMJS provides automated quality scoring:

Scores range from 0.00 to 1.00 and are recalculated daily. This lets agents prioritize higher-quality tools automatically.