Your MCPs Use Keyword Search. Documentation Needs Semantic Understanding.

AI That Actually
Knows Your Docs

GitHub MCP finds keywords. EmbeDocs understands MEANING. Perfect for finding what you need in vast documentation.

Star on GitHub
Without EmbeDocs ❌
You: How do I implement MongoDB vector search with Atlas?
GitHub MCP: Found 3 files with "vector" AND "search": - vector-search-tutorial.md - search-index.md - atlas-vector.md
With EmbeDocs ✅
You: How do I implement MongoDB vector search with Atlas?
EmbeDocs: Found 47 related concepts: vector indexes, $vectorSearch stage, kNN algorithms, embedding models, similarity scoring, hybrid search, performance tuning...
// Understands you need the COMPLETE solution

Why Keyword Search
Fails for Documentation

MCPs Miss Context

GitHub MCP finds files with exact keywords, missing related concepts

  • ✗ Need exact terms
  • ✗ Miss synonyms
  • ✗ No concept linking

Fragmented Results

Keyword search returns scattered pieces, not complete solutions

  • ✗ Partial information
  • ✗ Missing prerequisites
  • ✗ No related examples

Unknown Terminology

You often don't know the exact terms the docs use

  • ✗ "Sharding" vs "partitioning"
  • ✗ "Aggregation" vs "pipeline"
  • ✗ "Vector" vs "embedding"

Semantic Search
That Understands Meaning

Traditional search finds words. EmbeDocs understands concepts, relationships, and context

❌ GitHub MCP (Keyword)

// Query: "database performance"
// GitHub MCP searches for files containing both words

Found 3 files:
- database-performance.md
- performance.txt (mentions database once)
- troubleshooting-database-performance.md

// Missed: query optimization, indexing guides,
// connection pooling, caching, profiling...

✅ EmbeDocs (Semantic)

// Query: "database performance"
// Understands the CONCEPTS and RELATIONSHIPS

Found 127 semantically related docs:
- Query optimization techniques
- Index selection strategies  
- Read/write concern tuning
- Connection pool sizing
- Aggregation pipeline optimization
- Working set analysis
- Profiler interpretation
- Sharding for scale
- Caching patterns
- Hardware recommendations
1024 Dimensions
Deep Understanding
<100ms
Lightning Fast
50K+ Docs
Massive Scale
Git-Aware
Smart Updates

The Documentation Search Reality Check

Developers rarely know the exact terminology. Semantic search bridges this gap.

🤔 What You Search:

"How to make MongoDB queries faster"

Natural language, problem-focused

📚 What Docs Call It:

  • • Query Performance Optimization
  • • Index Selection Strategies
  • • Execution Plan Analysis
  • • Read Preference Configuration

🤔 What You Search:

"MongoDB not connecting"

Symptom-based, frustrated

📚 What Docs Call It:

  • • Connection String URI Format
  • • Network Access Configuration
  • • Authentication Mechanisms
  • • TLS/SSL Certificate Setup

🤔 What You Search:

"Store embeddings in database"

Task-oriented, learning

📚 What Docs Call It:

  • • Atlas Vector Search
  • • knnBeta Operator
  • • Dense Vector Storage
  • • Similarity Search Indexes

EmbeDocs understands both sides of this gap

Semantic search connects your natural language to technical documentation automatically

Setup in 60 Seconds
with Beautiful UI

EmbeDocs Beautiful Web Interface
1

Install EmbeDocs

One command to install globally

Installation Command
npm install -g embedocs-mcp
2

First Run (Auto Setup!)

🌐 Stunning web wizard opens automatically - visual setup with 2025 UI!

Launch Setup
embedocs
3

Start Indexing

✨ Web interface lets you pick popular repos or add custom ones - all automatic!

💡 Setup wizard handles everything - MongoDB Atlas + Voyage AI (both FREE), then start indexing with one click!

92%
Accuracy Score
Semantic relevance with MMR diversity
7.5x
Faster Search
Optimized vector search performance
50K+
Documents
Handle massive documentation sets

Upgrade from Keywords
to Semantic Understanding

Your AI deserves documentation search that understands meaning, not just matching words

Start Free NowView on GitHub
100% Free
Open Source
MIT License