paralel-search

Convert MCP tools into a typed Python SDK with async methods.

7|2|Updated Mar 3, 2026
One-click install
npx skills add https://github.com/manojbajaj95/mcp-skill --skill paralel-search
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: paralel-search
Source: https://github.com/manojbajaj95/mcp-skill/tree/main/skills/parallel_search
Command: npx skills add https://github.com/manojbajaj95/mcp-skill --skill paralel-search

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates direct programmatic access to MCP tools by turning each tool into a typed, async Python method, eliminating extra model prompts.

Core Features & Use Cases

  • Generated Typed Client: A Python App class with async methods for each MCP tool (web_search_preview, web_fetch).
  • Easy Authentication & Integration: OAuth support with token persistence to streamline subsequent runs.
  • Use Case: Build AI agents that programmatically search the web via MCP tools without embedding tool calls in the model.

Quick Start

Import the generated ParallelsearchApp and call web_search_preview with an objective and queries to get results directly from the MCP server.

Frequently Asked Questions about paralel-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I call MCP tools directly in Python without model back-and-forth?

Calling MCP tools directly in Python without model back-and-forth is done by using a generated typed SDK that converts each tool into an async method. This approach bypasses model prompts for fast, deterministic programmatic execution.

What is a typed Python async client for MCP tool invocation?

A typed Python async client for MCP tool invocation is a generated App class that provides async methods for each MCP tool. It manages OAuth authentication and JSON parsing to return deterministic search and fetch results directly.

How do I set up OAuth authentication for MCP web search calls in Python?

Setting up OAuth authentication for MCP web search calls in Python is handled natively by the generated client with token persistence. This streamlines subsequent runs by managing the OAuth flow automatically for your AI agents.

Can I use generated Python SDKs to build AI agents that search the web via MCP?

Yes, you can use generated Python SDKs to build AI agents that search the web via MCP tools. By importing the typed App class, agents invoke web_search_preview and web_fetch directly without embedding tool calls in the model.

Why use typed async methods instead of standard model tool-calling for MCP servers?

Using typed async methods instead of standard model tool-calling for MCP servers eliminates extra model prompts and ensures fast, deterministic execution. This direct programmatic access removes the latency and unpredictability of model back-and-forth.