06 / 10

Integrations & Ecosystem

GitHub · IDE · MCP · Cloud · Enterprise

How Claude connects to the tools and platforms you already use

API MCP SDK Cloud
OVERVIEW

The Integration Landscape

Claude AI Engine IDE Plugins VS Code / JetBrains / Cursor GitHub PRs / Actions / Issues Cloud Providers AWS / GCP / Azure MCP Servers Files / DB / APIs Automation Zapier / Slack / Make Enterprise SSO / VPC / Compliance Frameworks LangChain / LlamaIndex Productivity Notion / Docs / Email
GITHUB

Claude on GitHub

Claude integrates directly into GitHub workflows, acting as an AI-powered collaborator on your codebase.

PR Reviews

  • Automated code review on every pull request
  • Identifies bugs, security issues, and style violations
  • Suggests specific code improvements inline
  • Understands full repository context

Issue Triage

  • Auto-label and categorize new issues
  • Identify duplicate issues across the repo
  • Suggest assignees based on expertise
  • Draft initial responses for common questions

Code Suggestions

  • Generate implementation proposals from issues
  • Create draft PRs with suggested fixes
  • Refactor code on request via comments
  • Write and update tests automatically

Documentation

  • Auto-generate docs from code changes
  • Keep README files in sync with code
  • Generate changelog entries per release
  • Maintain API documentation
CI/CD

GitHub Actions with Claude

Example: Automated PR Review Workflow

name: Claude PR Review
on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Get diff
        run: |
          git diff origin/main...HEAD > diff.txt

      - name: Claude Review
        uses: anthropics/claude-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_KEY }}
          prompt: |
            Review this PR diff for bugs,
            security issues, and improvements.
          files: diff.txt
          max_tokens: 4096

      - name: Post Review Comment
        uses: actions/github-script@v7
        with:
          script: |
            github.rest.pulls.createReview({
              ...context.repo,
              pull_number: context.issue.number,
              body: process.env.CLAUDE_RESPONSE,
              event: 'COMMENT'
            })

Common Workflow Triggers

  • pull_request — review on open/update
  • issues — triage on creation
  • push — analyze on merge to main
  • schedule — periodic codebase audits
  • workflow_dispatch — manual triggers

Best Practices

  • Store API keys in GitHub Secrets
  • Set appropriate token limits
  • Use fetch-depth: 0 for full context
  • Rate-limit to avoid API quota issues
  • Cache responses for identical diffs
IDE

VS Code Extension

The Claude VS Code extension brings AI assistance directly into your editor with deep workspace integration.

Inline Suggestions

  • Context-aware code completions
  • Multi-line function generation
  • Smart import suggestions
  • Docstring auto-generation
  • Tab to accept, Esc to dismiss

Chat Panel

  • Side panel for conversations
  • Reference files with @file
  • Reference symbols with @symbol
  • Ask questions about your codebase
  • Generate code from descriptions

Commands

  • Explain Selection
  • Fix This Code
  • Add Tests
  • Refactor Selection
  • Generate Documentation

Setup

Install from VS Code Marketplace → Search "Claude" → Enter API key or sign in with Anthropic account → Configure model preferences in settings. Supports workspace-level and user-level configuration.

IDE

JetBrains Plugin

Claude support across the entire JetBrains ecosystem with IDE-native integration.

Supported IDEs

  • IntelliJ IDEA — Java, Kotlin, Scala
  • PyCharm — Python data science and web
  • WebStorm — JavaScript, TypeScript, React
  • Rider — C#, .NET development
  • CLion — C, C++, Rust
  • GoLand — Go development
  • PhpStorm — PHP and Laravel

IDE-Specific Advantages

  • Uses JetBrains' PSI for deep code understanding
  • Integrates with built-in refactoring tools
  • Respects IDE-level code style settings
  • Works with project-specific SDK configurations

Key Features

  • Tool window for chat-based interaction
  • Intention actions via Alt+Enter
  • Smart code generation from context
  • Test generation for selected classes
  • Commit message suggestions
  • Code review before commit

Installation

Settings → Plugins → Marketplace → Search "Claude" → Install → Restart IDE → Configure API key in Settings → Tools → Claude

IDE

Cursor IDE

Cursor is an AI-native code editor built on VS Code with deep Claude integration as a first-class model option.

How It Differs from VS Code

  • AI-first architecture — not a plugin but core
  • Codebase indexing — understands entire project
  • Composer — multi-file edit from description
  • Cmd+K inline editing — natural language edits
  • Tab completion — context-aware predictions
  • @-mentions — reference files, docs, web

Claude-Specific Features

  • Select Claude as preferred model in settings
  • Claude Sonnet for fast completions
  • Claude Opus for complex reasoning tasks
  • Maintains full VS Code extension compatibility
  • Supports custom system prompts per project
  • Privacy mode: code never stored on servers

When to Choose Cursor

Ideal for developers who want AI-assisted editing as a core workflow rather than an add-on. Best for multi-file refactors, rapid prototyping, and teams that want consistent AI-assisted development environments.

MCP

Model Context Protocol

MCP is an open standard that lets Claude connect to external data sources and tools through a unified protocol.

Host Application Claude Desktop / IDE / Custom App MCP Client JSON-RPC 2.0 over stdio / HTTP+SSE Filesystem Server read / write / search Database Server query / schema / migrate API Server REST / GraphQL / gRPC Local Files PostgreSQL / SQLite External Services

Why MCP Matters

Standardized protocol replacing ad-hoc tool integrations with a single open specification

Key Primitives

Tools (actions), Resources (data), Prompts (templates), Sampling (LLM requests)

Transport

Supports stdio for local servers and HTTP+SSE for remote servers

MCP

MCP Server Types

Filesystem Servers

  • Read and write local files securely
  • Directory listing and search
  • Configurable access boundaries
  • Watch for file changes

Database Servers

  • PostgreSQL, MySQL, SQLite connectors
  • Schema introspection and queries
  • Read-only mode for safety
  • Query result formatting

API Servers

  • GitHub, GitLab, Jira, Linear
  • Slack, Discord, email services
  • Cloud provider APIs (AWS, GCP)
  • Custom REST/GraphQL endpoints

Community Servers

  • Growing open-source ecosystem
  • Puppeteer for browser automation
  • Google Drive, Notion, Obsidian
  • Docker, Kubernetes management

Server Discovery

Find servers at github.com/modelcontextprotocol/servers — the official registry of reference and community MCP server implementations. Configuration is managed through claude_desktop_config.json or project-level .mcp.json files.

MCP

Building Custom MCP Servers

TypeScript MCP Server Example

import { Server } from "@modelcontextprotocol/sdk/server";
import { StdioTransport }
  from "@modelcontextprotocol/sdk/server/stdio";

const server = new Server({
  name: "my-custom-server",
  version: "1.0.0",
});

// Define a tool
server.setRequestHandler(
  "tools/list",
  async () => ({
    tools: [{
      name: "search_docs",
      description: "Search internal documentation",
      inputSchema: {
        type: "object",
        properties: {
          query: {
            type: "string",
            description: "Search query"
          }
        },
        required: ["query"]
      }
    }]
  })
);

// Handle tool calls
server.setRequestHandler(
  "tools/call",
  async (request) => {
    if (request.params.name === "search_docs") {
      const results = await searchDocs(
        request.params.arguments.query
      );
      return {
        content: [{
          type: "text",
          text: JSON.stringify(results)
        }]
      };
    }
  }
);

const transport = new StdioTransport();
await server.connect(transport);

Available SDKs

  • TypeScript — @modelcontextprotocol/sdk
  • Python — mcp Python package
  • Kotlin — JVM-based server SDK
  • C# — .NET SDK

Server Architecture

  • Define tools Claude can invoke
  • Expose resources as read-only data
  • Provide prompts as reusable templates
  • Handle lifecycle and error management

Deployment Options

  • Local — stdio transport, runs as subprocess
  • Remote — HTTP+SSE, deploy anywhere
  • Container — Docker for isolation
  • Serverless — AWS Lambda, Cloud Functions
CLOUD

AWS Bedrock

Access Claude through Amazon Bedrock — a fully managed service that keeps data within your AWS environment.

Bedrock API Example

import boto3
import json

client = boto3.client(
    "bedrock-runtime",
    region_name="us-east-1"
)

response = client.invoke_model(
    modelId="anthropic.claude-sonnet-4-20250514",
    body=json.dumps({
        "anthropic_version": "bedrock-2023-05-31",
        "messages": [{
            "role": "user",
            "content": "Explain microservices."
        }],
        "max_tokens": 1024,
        "temperature": 0.7
    })
)

result = json.loads(response["body"].read())
print(result["content"][0]["text"])

Key Advantages

  • Data stays within your AWS VPC
  • Use existing IAM roles and policies
  • Pay per token via AWS billing
  • No separate API key management
  • SOC 2 / HIPAA compliance via AWS

Available Models

  • Claude Opus — most capable
  • Claude Sonnet — balanced performance
  • Claude Haiku — fastest, lowest cost

Setup Requirements

Enable model access in Bedrock console → Configure IAM permissions → Select region with model availability

CLOUD

Google Cloud Vertex AI

Claude is available on Google Cloud's Vertex AI platform, bringing enterprise AI capabilities to the GCP ecosystem.

Vertex AI Example

from anthropic import AnthropicVertex

client = AnthropicVertex(
    project_id="my-gcp-project",
    region="us-east5"
)

response = client.messages.create(
    model="claude-sonnet-4@20250514",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "Design a REST API schema."
    }]
)

print(response.content[0].text)

GCP Integration Benefits

  • Use existing GCP service accounts
  • Integrate with BigQuery for data analysis
  • Connect to Cloud Storage pipelines
  • Unified billing through GCP console
  • Regional data residency compliance

Configuration

  • Enable Vertex AI API in GCP console
  • Enable Claude models in Model Garden
  • Set up authentication via ADC or service account
  • Use the anthropic[vertex] Python SDK

Supported Regions

us-east5, europe-west1, asia-southeast1 — check availability for latest regions

NO-CODE

Zapier & Make

Connect Claude to thousands of apps without writing code using visual automation platforms.

Zapier Integration

  • Official Claude action in Zapier
  • Trigger: new email, form, Slack message
  • Action: send to Claude for processing
  • Output: route result to any connected app
  • Supports multi-step Zaps with Claude

Popular Zap Templates

  • Summarize emails → post to Slack
  • Analyze form responses → update spreadsheet
  • Translate support tickets → route to team
  • Generate social posts from blog content

Make (Integromat)

  • Claude module with full API access
  • Visual scenario builder with branching
  • Iterator and aggregator support
  • Process batches of data through Claude
  • Error handling and retry logic built in

Use Cases

  • Customer support — auto-draft responses
  • Content ops — repurpose across channels
  • Data enrichment — classify and tag records
  • Reporting — generate summaries from data
CHAT

Slack Integration

Bring Claude into your team's Slack workspace for real-time AI assistance in conversations.

How It Works

  • Add Claude as a Slack app
  • Mention @Claude in any channel
  • DM Claude for private conversations
  • Claude reads thread context
  • Responds within Slack natively

Team Use Cases

  • Answer technical questions in channels
  • Summarize long threads on demand
  • Draft responses for customer queries
  • Brainstorm ideas in group channels
  • Translate messages across languages

Admin Controls

  • Restrict to specific channels
  • Set per-user usage limits
  • Configure data retention policies
  • Audit log for all interactions
  • SSO integration for access control

Building Custom Slack Bots with Claude

Use the Slack Bolt SDK combined with the Anthropic API to build custom bots that handle specific workflows: ticket triage, standup summaries, onboarding assistants, and more. Deploy on AWS Lambda or your own infrastructure.

PRODUCTIVITY

Notion & Productivity Tools

Claude integrates with productivity platforms to enhance document creation, knowledge management, and workflows.

Notion AI (powered by Claude)

  • Built-in AI writing and editing assistance
  • Summarize pages and databases
  • Generate content from templates
  • Q&A over your workspace knowledge
  • Translate content across languages
  • Extract action items from meeting notes

Document Assistants

  • Google Docs — drafting and editing add-ons
  • Confluence — technical doc generation
  • Obsidian — personal knowledge assistant via MCP
  • Coda — AI-powered formulas and summaries

Knowledge Management

  • Auto-tag and categorize documents
  • Generate summaries for search indexes
  • Link related documents automatically
  • Create glossaries from technical content

Email & Calendar

  • Draft email responses with context
  • Summarize email threads
  • Generate meeting agendas from topics
  • Create follow-up action items
ENTERPRISE

Enterprise Integration Patterns

Single Sign-On (SSO)

  • SAML 2.0 and OIDC support
  • Okta, Azure AD, Google Workspace
  • Role-based access control (RBAC)
  • Automated user provisioning via SCIM

Data Residency

  • US and EU data regions
  • GDPR-compliant data processing
  • Data never used for model training
  • Configurable data retention periods

VPC & Network Security

  • AWS PrivateLink for Bedrock access
  • VPC Service Controls on GCP
  • IP allowlisting for API access
  • No data traverses public internet

Compliance & Governance

  • SOC 2 Type II certified
  • HIPAA eligible (via cloud partners)
  • Audit logs for all API interactions
  • Content filtering and safety controls

Deployment Architecture

Enterprise customers typically deploy via cloud partners (Bedrock / Vertex) for data sovereignty, use API gateways for rate limiting and monitoring, and maintain centralized key management through their secrets manager (AWS Secrets Manager, HashiCorp Vault).

FRAMEWORKS

LangChain & LlamaIndex

Use Claude with popular AI orchestration frameworks for building complex, multi-step applications.

LangChain with Claude

from langchain_anthropic import ChatAnthropic
from langchain.agents import create_tool_calling_agent
from langchain.tools import tool

llm = ChatAnthropic(
    model="claude-sonnet-4-20250514",
    temperature=0
)

@tool
def search_db(query: str) -> str:
    """Search the product database."""
    return db.search(query)

agent = create_tool_calling_agent(
    llm=llm,
    tools=[search_db],
    prompt=prompt_template
)

result = agent.invoke({
    "input": "Find top-selling products"
})

LlamaIndex with Claude

from llama_index.llms.anthropic import Anthropic
from llama_index.core import (
    VectorStoreIndex,
    SimpleDirectoryReader
)

llm = Anthropic(
    model="claude-sonnet-4-20250514"
)

# Build index from documents
documents = SimpleDirectoryReader(
    "./data"
).load_data()

index = VectorStoreIndex.from_documents(
    documents
)

# Query with Claude as the LLM
query_engine = index.as_query_engine(
    llm=llm
)

response = query_engine.query(
    "What are our return policies?"
)

LangChain Strengths

Chains, agents, tool calling, memory management, output parsing, and a large ecosystem of integrations

LlamaIndex Strengths

RAG pipelines, document indexing, vector stores, knowledge graph queries, and structured data extraction

API

Custom Application Integration

Embed Claude directly into your applications using official SDKs and the Messages API.

Official SDKs

// TypeScript SDK
import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic();

// Streaming response
const stream = await client.messages.stream({
  model: "claude-sonnet-4-20250514",
  max_tokens: 1024,
  system: "You are a helpful assistant.",
  messages: [
    { role: "user", content: userInput }
  ]
});

for await (const event of stream) {
  if (event.type === "content_block_delta") {
    process.stdout.write(
      event.delta.text
    );
  }
}

SDK Availability

  • Pythonanthropic
  • TypeScript@anthropic-ai/sdk
  • Javaanthropic-sdk-java
  • Go, Rust — community SDKs
  • REST — direct HTTP for any language

Best Practices

  • Use streaming for responsive UIs
  • Implement retry logic with exponential backoff
  • Cache responses for identical queries
  • Set appropriate timeouts per use case
  • Monitor token usage and costs

Common Patterns

Chat interfaces, document analysis pipelines, content generation systems, code assistants, and multi-agent workflows

SECURITY

Integration Security

Securing your Claude integrations requires attention to key management, data handling, and compliance.

API Key Management

  • Never commit keys to version control
  • Use environment variables or secrets managers
  • Rotate keys on a regular schedule
  • Create separate keys per environment
  • Set per-key spending limits
  • Use workspace-scoped keys in teams

Data Handling

  • Sanitize PII before sending to API
  • Implement input validation and filtering
  • Use data classification for prompts
  • Log requests for audit trails
  • Respect data residency requirements
  • Encrypt data in transit (TLS 1.2+)

Access Controls

  • Principle of least privilege for API access
  • Rate limiting per user and per endpoint
  • Monitor for anomalous usage patterns
  • Implement request signing where possible
  • Use OAuth 2.0 for user-facing integrations
  • Regular security audits of integrations

Compliance Considerations

SOC 2 — audit logging and access controls. GDPR — data processing agreements and right to deletion. HIPAA — use Bedrock/Vertex with BAA. PCI-DSS — never send cardholder data to the API. Always review Anthropic's usage policies for your specific compliance requirements.

GUIDE

Choosing the Right Integration

Match your use case to the optimal integration approach.

Use Case Recommended Integration Complexity Best For
Code assistance in editor VS Code / Cursor / JetBrains Low Individual developers
Automated PR reviews GitHub Actions Low-Medium Development teams
No-code automation Zapier / Make Low Non-technical users
Team communication Slack integration Low Collaborative teams
Custom tools for Claude MCP servers Medium Power users, developers
RAG & document Q&A LangChain / LlamaIndex Medium AI application builders
Enterprise data sovereignty AWS Bedrock / Vertex AI Medium-High Enterprise, regulated industries
Fully custom application Direct API / SDK High Product teams, SaaS builders

Decision Framework

Start with the simplest integration that meets your needs. Move to more complex options only when you need greater control, customization, or compliance guarantees. Many teams use multiple integrations simultaneously.

SUMMARY

Integrations & Ecosystem

Developer Tools

  • VS Code, JetBrains, Cursor — AI in your editor
  • GitHub Actions — automated code review and CI
  • LangChain, LlamaIndex — orchestration frameworks
  • SDKs — Python, TypeScript, Java, and REST

Model Context Protocol

  • Open standard for tool and data integration
  • Growing ecosystem of community servers
  • Build custom servers with official SDKs
  • Filesystem, database, API connectors

Cloud & Enterprise

  • AWS Bedrock — VPC, IAM, compliance
  • Google Vertex AI — GCP-native integration
  • SSO, RBAC, data residency for enterprise
  • SOC 2, GDPR, HIPAA compliance paths

Automation & Productivity

  • Zapier, Make — no-code workflows
  • Slack — team AI assistant
  • Notion — knowledge management
  • Start simple, scale as needed

Next: 07 — Safety & Alignment