Build with Context

For developers: run Context agents from your own code with the Agent API and SDKs, and measure them with the Evals API and CLI.

Agent API and SDKsStart a session with one of your organization’s agents, stream its events, approve its tool calls and read the files it produced, with Anthropic’s Python, TypeScript and .NET SDKs or plain HTTP.Evals API and CLIRead evaluation results and traces over HTTP, and build and run evaluations from the command line.

Agent API and SDKs

• Quickstart: upload a CSV file, ask an agent a question about it and read the answer in about five minutes.
• SDKs: point Anthropic’s Python, TypeScript or .NET SDK at your deployment.
• Choose a key: a personal Task API key or a service account key.
• Coming from Anthropic’s API: what works unchanged, what to change, and what each unsupported call returns.
• Playbooks: end-to-end flows, tested against a deployment.
• API reference: every endpoint, generated from the OpenAPI file.

Evals API and CLI

• Quickstart, Compare models, Read results, Monitors.
• CLI reference and API reference.

API keys, versioning, errors and limits for the Evals API are under Overview.