Add Exa AI MCP Server - Semantic Search Integration #359
Conversation
Implements 5 atomic tools for intelligent web search and content discovery: ✅ search_web_semantic - Neural-powered web search with natural language understanding ✅ extract_page_content - Clean text extraction from any web page ✅ find_similar_pages - Content similarity discovery using AI matching ✅ search_recent_content - Time-filtered search for current information ✅ search_by_example_text - Text-based semantic similarity search Key Features: - Semantic understanding vs keyword matching using Exa's neural networks - Comprehensive error handling with clear, actionable messages - Production-ready architecture with detailed logging - AI-friendly tool descriptions optimized for LLM consumption - Robust input validation and sanitization Technical Implementation: - Full MCP protocol compliance with proper tool schemas - Atomic tool design following Klavis AI guidelines - Clean code architecture with separation of concerns - Professional documentation and setup instructions - Comprehensive testing suite with API validation Integration: - Works seamlessly with Claude Desktop and other MCP clients - Includes wrapper scripts for reliable deployment - Environment variable configuration for API credentials - Health checks and connection validation Value Proposition: - Transforms traditional keyword search into intelligent semantic discovery - Enables AI assistants to find contextually relevant, high-quality information - Provides professional-grade search capabilities for the Klavis AI ecosystem Ready for production deployment and integration into Klavis AI platform.
|
Tharun D Chowdary seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account. You have signed the CLA already but the status is still pending? Let us recheck it. |
OverviewThis PR adds a comprehensive MCP server that integrates Exa AI's semantic search capabilities into the Klavis AI ecosystem. ImplementationFollowing the Klavis AI guidelines for building custom MCP servers, this implementation provides 5 atomic tools that transform Exa AI's neural search technology into AI-friendly capabilities. Tools Provided
Klavis AI Guidelines Compliance✅ Atomicity - Each tool performs one specific job perfectly Technical Highlights
Testing & Validation
Proof of CorrectnessIncludes video demonstrations showing:
Value PropositionThis integration brings Exa AI's cutting-edge semantic search capabilities to the Klavis AI ecosystem, enabling AI assistants to find contextually relevant, high-quality information rather than just keyword matches. This represents a significant upgrade from traditional search APIs. Files Addedmcp_servers/exa/ |
There was a problem hiding this comment.
Thanks for the contri, but we can't merge it unless you sign cla.
Can you provide us proof that its working, like photo or video
Also some changes are proposed in comments, would like you to see that,
I would like you to follow this architecture, - https://github.com/Klavis-AI/klavis/tree/main/mcp_servers/hubspot
also can you add a doc for it, like these - https://github.com/Klavis-AI/klavis/tree/main/docs/documentation/mcp-server
| conda activate base | ||
| export EXA_API_KEY="29966140-48f9-4786-8c70-fd539db903be" | ||
| cd /Users/tharundchowdary/exa-mcp-server-klavis | ||
| python server.py |
There was a problem hiding this comment.
bro need a dockerfile, not this,
and also you exposed your key, so better to revoke your key
docker like this - https://github.com/Klavis-AI/klavis/blob/main/mcp_servers/hubspot/Dockerfile
| high-quality Model Context Protocol (MCP) server that provides AI-powered web search capabilities through Exa's sophisticated search engine. This server exposes Exa's full suite of search tools, enabling AI agents to perform semantic search, content retrieval, similarity finding, direct question answering, and comprehensive research. | ||
|
|
||
| ## Purpose | ||
|
|
There was a problem hiding this comment.
can you add what tools are there in your mcp in readme file
| # Exa AI MCP Server Dependencies | ||
| # Core MCP and Exa requirements | ||
| exa-py>=1.0.0 | ||
| mcp>=1.0.0 |
| pytest>=7.0.0 | ||
| pytest-asyncio>=0.21.0 |
| @@ -0,0 +1,284 @@ | |||
| #!/usr/bin/env python3 | |||
There was a problem hiding this comment.
remove this file
| @@ -1,627 +1,424 @@ | |||
| import contextlib | |||
| import base64 | |||
| #!/usr/bin/env python3 | |||
There was a problem hiding this comment.
instead of putting everything in server.py, can you make server.py like this one
https://github.com/Klavis-AI/klavis/blob/main/mcp_servers/hubspot/server.py
and add tools in seperate folder like this one-
https://github.com/Klavis-AI/klavis/tree/main/mcp_servers/hubspot/tools
Implements 5 atomic tools for intelligent web search and content discovery:
✅ search_web_semantic - Neural-powered web search with natural language understanding
✅ extract_page_content - Clean text extraction from any web page
✅ find_similar_pages - Content similarity discovery using AI matching
✅ search_recent_content - Time-filtered search for current information
✅ search_by_example_text - Text-based semantic similarity search
Key Features:
Technical Implementation:
Integration:
Value Proposition:
Ready for production deployment and integration into Klavis AI platform.
Description
Related issue
Type of change
How has this been tested?
(Add screenshots or recordings here if applicable.)
Checklist