Building effective agents
Listed inAgentsBuilding AI Featureson
Anthropic on the difference between workflows and agents, five composable patterns, and why most teams should start with the simplest thing that works.
Shipping model-backed UI: streaming, tool calls, and the loading and error states nobody designs first.
9 articles
Listed inAgentsBuilding AI Featureson
Anthropic on the difference between workflows and agents, five composable patterns, and why most teams should start with the simplest thing that works.
Listed inModel Context ProtocolBuilding AI Featureson
The normative source: lifecycle, capability negotiation, transports, and the exact shape of tools, resources, and prompts.
Listed inGeminiBuilding AI Featureson
Google’s API, its multimodal inputs, and the long-context behaviour to design around.
Listed inOpenAIBuilding AI Featureson
The OpenAI API surface — function calling, structured outputs, and the JavaScript SDK.
Listed inAnthropicBuilding AI Featureson
The Claude API from a frontend perspective: messages, tool use, and streaming.
Listed inModel Context ProtocolBuilding AI Featureson
A shared protocol for exposing tools and data to models, instead of N×M bespoke integrations.
Listed inAgentsBuilding AI Featureson
Tool-calling loops in a product surface: turn limits, confirmation steps, and showing the work.
Listed inStreamed ResponsesBuilding AI Featureson
Rendering tokens as they arrive over SSE — backpressure, cancellation, and partial markdown.
Listed inImplementing AIBuilding AI Featureson
Where the API call belongs, how keys stay server-side, and the UX of a slow, non-deterministic response.