Choosing a Claude Model
The Opus/Sonnet/Haiku tradeoff and how to pick — without guessing.
Tokens, Context & Pricing
Count tokens correctly, max_tokens vs context window, and estimating cost.
Your First API Call
Install the SDK, authenticate, and make a real call to Claude in Python, TypeScript, or cURL.
Streaming & Multi-Turn Conversations
Stream responses and manage history on a stateless API.
Tool Use / Function Calling
Define tools, run the agentic loop, and handle results.
Vision, PDF & File Input
Send images and documents, and reuse uploads by file_id.
Structured Output
Get schema-conforming JSON you can rely on — not prompt-and-pray.
Extended Thinking & Effort
Adaptive reasoning and the effort control — choosing depth by workload.
Prompt Caching & Cost Optimization
Reuse a stable prompt prefix to cut cost and latency — and the gotcha that breaks it.
Building Agents on the API
Single call vs workflow vs custom agent — the decision test and loop design.
Managed Agents
Anthropic-hosted agent loop: sessions, environments, memory, scheduling.
Cowork & Agent Teams
The agentic desktop workspace and multi-agent collaboration as products.
MCP & Connecting to Tools
Declare MCP servers on requests, and where MCP fits vs other tool types.
Safety, Refusals & Fallbacks
Handle refusals gracefully and distinguish them from classifier blocks.
Errors, Rate Limits & Reliability
The error map, retries with backoff, rate-limit tiers, and model migration.