AG-Bureau/mcp-search

Overview

A Python-based Model Context Protocol server for web research. It performs web searches, reads pages, fetches images, captures screenshots, and runs multi-wave research. Each answer includes a confidence score.

Best for

Best for
Python developers building MCP-based research agents.

Use cases

• Add web search and page reading to an MCP client
• Run iterative multi-wave research queries
• Collect screenshots and images as part of a research pipeline

How to use

Tools exposed

• web_search
• web_read
• web_image_search
• web_screenshot
• web_deep_search
• unresponsive_engines
• engines_irrelevant
• arguments_adjusted
• SEARXNG_SECRET
• LLM_API_KEY
• LLM_API_BASE
• LLM_MODEL_TEXT
• LLM_MODEL_VISION
• LLM_DISABLE_THINKING
• READ_CONTACT
• READ_LANGUAGES

Tested with

Continue, ChatGPT

Notes

A Python-based Model Context Protocol server for web research. It performs web searches, reads pages, fetches images, captures screenshots, and runs multi-wave research. Each answer includes a confidence score.

0 stars on GitHub. Last updated 2026-09-09. Licensed AGPL-3.0.

Use cases

• Add web search and page reading to an MCP client
• Run iterative multi-wave research queries
• Collect screenshots and images as part of a research pipeline

Pros

• Returns a confidence score with each answer
• Covers search, reading, images, and screenshots in one tool
• Simple Python setup for MCP environments

Cons

• No community adoption yet (zero stars)
• Requires an MCP client to be useful
• Confidence scores depend on source quality

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.