Guides Overview¶
These guides provide deep dives into specific features and patterns for building IVCAP lambda services.
Getting Started¶
If you're new to the SDK, start here:
- Installation — Set up the SDK
- Quick Start — 5-minute introduction
- Your First Lambda Service — Build a real service
Core Concepts¶
Tool Functions¶
Learn how to define and register tools:
- Request/Result schemas with Pydantic
- The @ivcap_lambda decorator and ToolOptions
- Accessing JobContext and the FastAPI Request
- Async tool functions
- Try-later semantics for long-running calls
Read this if: You want to understand the fundamentals of building lambda tool handlers.
Working with Artifacts¶
Upload and download IVCAP artifacts: - Downloading input artifacts - Uploading results - Handling file streams - Working with collections
Read this if: Your tool processes or generates files.
MCP & Agent Integration¶
Expose tools to AI agent frameworks, covering both directions:
- Adding MCP (--with-mcp/--with-mcp-stdio) to an existing IVCAP lambda service
- Building an MCP server first, with IVCAP deployment as an optional bonus
- Connecting MCP clients (Claude Desktop, Cline, MCP Inspector)
- Tool description endpoints
Read this if: You want your tools to be directly callable by AI agents/MCP hosts via the Model Context Protocol — whether that's your primary goal or an add-on to a REST service.
Advanced Topics¶
Observability & Logging¶
Monitor your services:
- Structured logging with getLogger and logging_init
- Progress events and step reporting
- OpenObserve integration
- OpenTelemetry tracing with FastAPI
Read this if: You want comprehensive monitoring and debugging.
Error Handling¶
Build robust services: - Exception handling patterns - Reporting errors through the event system - HTTP status codes returned - Graceful degradation
Read this if: You need production-grade error handling.
Deployment¶
Deploy to production: - Docker containerization - Environment configuration - Service registration with IVCAP - Kubernetes patterns
Read this if: You're ready to deploy your service.
Best Practices¶
Pro tips and patterns: - Code organization - Testing strategies - Performance optimization - Common pitfalls
Read this if: You want to write production-quality code.
Learning Path¶
Installation
↓
Quick Start
↓
Your First Lambda Service
↓
Tool Functions
├→ Artifacts (if needed)
├→ MCP & Agent Integration (if needed)
└→ Observability (if needed)
↓
Error Handling
↓
Deployment
↓
Best Practices
Choose Your Own Path¶
- I want to expose a tool to AI agents: Tool Functions → MCP & Agent Integration
- I need to process files: Artifacts
- I want to monitor my service: Observability
- I need robust error handling: Error Handling
- I'm ready for production: Deployment → Best Practices
See Also¶
- API Reference — Complete API documentation
- Examples — Working code samples
- Reference — Configuration and schemas