Context7 Language Integration

Context7 Language Integration

Quick Reference

What It Does

This Skill provides enterprise-grade Context7 MCP integration patterns for accessing the latest documentation across all programming languages and frameworks. It centralizes common patterns that Language Skills reuse to reduce code duplication and token overhead.

Core Capabilities:

• Resolve library names to Context7-compatible IDs in real-time
• Fetch latest documentation (automatically updated)
• Intelligent caching with TTL-based invalidation
• Progressive token disclosure (1000-10000 tokens)
• Graceful error handling and fallback strategies

When to Use

Use this Skill when:

• Building language-specific Skills that need external documentation
• Integrating latest framework/​library documentation into guides
• Creating up-to-date API reference materials
• Reducing token consumption through centralized caching
• Implementing documentation access in specialized agents

Core Concepts

1. Library ID Format

• Format: /​org/​project or /​org/​project/​version
• Examples: /​tiangolo/​fastapi, /​tiangolo/​fastapi/​v0.115.0
• Returned by resolve-library-id, used by get-library-docs

2. Two-Step Integration Pattern

• Step 1: resolve-library-id("library-name") → get canonical ID
• Step 2: get-library-docs(canonical-id) → fetch documentation
• Separation allows caching at both layers

3. Progressive Disclosure

• Token budget: 1000-10000 tokens per request
• Default: 5000 tokens (balanced)
• Specific topic focus reduces token consumption
• Multiple requests accumulate toward session limit

4. Error Handling Strategy

• Try primary library name
• Fallback to alternative names (case variations)
• Return manual documentation link as last resort
• Log failures for debugging and optimization

5. Caching Strategy

• Cache key: {library_id}:{topic}:{tokens}
• TTL: 30 days (stable versions)
• TTL: 7 days (beta/​latest versions)
• Storage: .moai/​cache/​context7/
• Monitoring: Track hit/​miss rates

Quick Start Example

# Step 1: Resolve library name to Context7 ID
from mcp__context7__resolve_library_id import resolve_library_id
from mcp__context7__get_library_docs import get_library_docs

library_name = "fastapi"
library_id = resolve_library_id(library_name)
# Returns: /​tiangolo/​fastapi or /​tiangolo/​fastapi/​v0.115.0

# Step 2: Get documentation with topic focus
docs = get_library_docs(
    context7_compatible_library_id=library_id,
    tokens=3000,
    topic="routing"
)

# Step 3: Use in workflow
print(f"Documentation for {library_name}:\n{docs[:500]}")

---

Implementation

Step 1: Library Resolution Pattern

Purpose: Convert user-friendly library names to canonical Context7 IDs

Pattern Design:

Input: User-friendly name (e.g., "fastapi")
Process:
  1. Validate input (non-empty, reasonable length)
  2. Query Context7: resolve-library-id MCP
  3. Return canonical ID: /​org/​project/​version
  4. Cache for 30 days
Output: Context7-compatible library ID

Implementation Notes:

• Always validate library name before query (non-empty, <50 chars)
• Implement retry logic (3 attempts max)
• Cache results with TTL (30 days stable, 7 days beta)
• Log failed resolutions for optimization
• Return structured error for missing libraries

Error Cases:

• Library not found → Try alternative names
• Network timeout → Use cached version
• Malformed name → Return clear error message

Step 2: Documentation Fetching Pattern

Purpose: Get specific documentation topics to optimize token usage

Pattern Design:

Input: Library ID + topic + token limit
Process:
  1. Check cache for {library_id}:{topic}:{tokens}
  2. If cache hit and valid (TTL), return cached
  3. If cache miss, query Context7 API
  4. Store result in cache
Output: Markdown-formatted documentation

Token Optimization Strategy:

• Default: 5000 tokens (balanced documentation)
• Specific topic: 3000 tokens (focused API docs)
• Summary only: 1000 tokens (quick reference)
• Comprehensive: 10000 tokens (full reference)

Progressive Disclosure Levels:

• Level 1 (1000 tokens): Core API only
• Level 2 (3000 tokens): API + examples
• Level 3 (5000 tokens): Full documentation
• Level 4 (10000 tokens): Full docs + advanced

Step 3: Error Handling Pattern

Case 1: Library Not Found

Error: resolve-library-id fails (LibraryNotFoundError)
Recovery:
  1. Try alternative names: "FastAPI" → "fast-api"
  2. Search library registry manually
  3. Fallback: Provide manual documentation link
  4. Log for optimization

Case 2: Documentation Unavailable

Error: get-library-docs returns empty (DocumentationNotFoundError)
Recovery:
  1. Retry with broader topic ("routing" → "api")
  2. Fetch summary instead (1000 tokens)
  3. Use cached version from previous fetch
  4. Provide manual link

Case 3: Token Limit Exceeded

Error: Requested tokens exceed available (TokenLimitExceededError)
Recovery:
  1. Reduce token count automatically
  2. Retry with minimum viable tokens (1000)
  3. Split request into multiple focused queries
  4. Implement budget management across requests

Step 4: Language Skills Integration Pattern

Integration Points: Each Language Skill should include:

## External Documentation Access

For up-to-date library documentation:
Skill("moai-context7-lang-integration")

**Available libraries**:
- FastAPI: `/​tiangolo/​fastapi`
- Django: `/​django/​django`
- Pydantic: `/​pydantic/​pydantic`
- SQLAlchemy: `/​sqlalchemy/​sqlalchemy`

Usage in Implementation:

def get_fastapi_routing_guide(tokens: int = 3000) -> str:
    # Use Context7 integration
    library_id = "/​tiangolo/​fastapi"
    docs = get_library_docs(
        context7_compatible_library_id=library_id,
        topic="routing",
        tokens=tokens
    )
    return docs

Step 5: Caching Implementation Pattern

Cache Architecture:

Cache Structure:
  Key: {library_id}:{topic}:{tokens}
  Value: {content, timestamp, version}
  Storage: .moai/​cache/​context7/
  Index: .moai/​cache/​context7/​index.json

Caching Rules:

• TTL: 30 days for stable versions (v1.0, v2.0)
• TTL: 7 days for latest/​beta (main, beta, rc)
• TTL: 1 day for development branches
• Hit rate tracking (monitor for optimization)

Cache Invalidation Strategy:

• Auto-invalidate on version mismatch
• Manual invalidation option (--clear-cache)
• Smart invalidation (version changed)

Implementation Checklist:

• Create cache directory if missing
• Implement TTL validation on read
• Log cache hits/​misses for metrics
• Implement cache cleanup (retention policy)
• Monitor for stale entries

---

Advanced

Multi-Library Integration Pattern

Use Case: Build comprehensive tech stack documentation

def get_tech_stack_docs(stack_name: str) -> dict:
    """Fetch documentation for entire tech stack"""

    stacks = {
        "modern-python": {
            "fastapi": "/​tiangolo/​fastapi",
            "pydantic": "/​pydantic/​pydantic",
            "sqlalchemy": "/​sqlalchemy/​sqlalchemy"
        },
        "modern-js": {
            "nextjs": "/​vercel/​next.js",
            "react": "/​facebook/​react",
            "typescript": "/​microsoft/​TypeScript"
        }
    }

    libraries = stacks.get(stack_name, {})
    results = {}

    for lib_name, lib_id in libraries.items():
        docs = get_library_docs(
            context7_compatible_library_id=lib_id,
            tokens=2000  # Shared token budget
        )
        results[lib_name] = docs

    return results

Parallel Request Pattern

Benefits: Reduce total fetch time for multiple libraries

import asyncio

async def fetch_library_async(lib_id: str, topic: str) -> str:
    # Implementation depends on async Context7 client
    return get_library_docs(
        context7_compatible_library_id=lib_id,
        topic=topic,
        tokens=3000
    )

async def fetch_multiple(libraries: list) -> dict:
    tasks = [
        fetch_library_async(lib['id'], lib['topic'])
        for lib in libraries
    ]
    results = await asyncio.gather(*tasks, return_exceptions=True)
    return dict(zip([l['name'] for l in libraries], results))

Token Budget Management Pattern

Strategy: Allocate and track tokens across multiple requests

class TokenBudgetManager:
    def __init__(self, total_budget: int = 20000):
        self.total_budget = total_budget
        self.allocated = {}
        self.used = {}

    def allocate(self, skill_name: str, percentage: int) -> int:
        tokens = int(self.total_budget * percentage / 100)
        self.allocated[skill_name] = tokens
        self.used[skill_name] = 0
        return tokens

    def get_available(self, skill_name: str) -> int:
        return self.allocated.get(skill_name, 0) - self.used.get(skill_name, 0)

    def use_tokens(self, skill_name: str, amount: int):
        available = self.get_available(skill_name)
        if amount > available:
            raise TokenBudgetExceededError(
                f"{skill_name} budget exceeded: {amount} > {available}"
            )
        self.used[skill_name] = self.used.get(skill_name, 0) + amount

Performance Optimization Pattern

Techniques:

1. Lazy Loading: Fetch only when explicitly needed

   docs = None  # Don't fetch until requested
   if user_needs_full_docs():
       docs = get_library_docs(...)
   

2. Partial Content: Request specific topic instead of full docs

   # Good: Fetch only routing documentation
   docs = get_library_docs(topic="routing", tokens=1000)
   # Bad: Fetch entire documentation
   docs = get_library_docs(tokens=10000)
   

3. Compression: Gzip cached documentation

   import gzip
   compressed = gzip.compress(docs.encode())  # ~60% reduction
   

4. Indexing: Create searchable index of cached docs

   index = {
       "routing": {offset: 0, length: 5000},
       "async": {offset: 5000, length: 3000},
   }
   

5. Version Pinning: Cache specific versions longer

   library_id = "/​tiangolo/​fastapi/​v0.115.0"  # Pinned version
   # Cached for 90 days instead of 30
   

Custom Library Mapping Pattern

Purpose: Map language-specific names to canonical Context7 IDs

LIBRARY_MAPPINGS = {
    "python": {
        "fastapi": "/​tiangolo/​fastapi",
        "FastAPI": "/​tiangolo/​fastapi",
        "fast-api": "/​tiangolo/​fastapi",
        "django": "/​django/​django",
        "Django": "/​django/​django",
        "pydantic": "/​pydantic/​pydantic",
    },
    "javascript": {
        "next.js": "/​vercel/​next.js",
        "nextjs": "/​vercel/​next.js",
        "NextJS": "/​vercel/​next.js",
        "react": "/​facebook/​react",
        "React": "/​facebook/​react",
    },
    "go": {
        "gin": "/​gin-gonic/​gin",
        "Gin": "/​gin-gonic/​gin",
        "gorm": "/​go-gorm/​gorm",
        "GORM": "/​go-gorm/​gorm",
    }
}

def resolve_with_mapping(language: str, library_name: str) -> str:
    mapping = LIBRARY_MAPPINGS.get(language, {})
    canonical_id = mapping.get(library_name)
    if canonical_id:
        return canonical_id
    # Fallback to MCP resolver
    return resolve_library_id(library_name)