
Claude Agent SDK
β Text source-checked Nov 22, 2026Build production AI agents with Claude Code as a library
Source-checked against the vendor's site Β· rebuilt from the vendor's own pagesProfile updated Link checked How we rate
In short
The Claude Agent SDK lets you build production AI agents in Python or TypeScript using Claude Code's core tools and loop. It supports file manipulation, command execution, and MCP integrations.
The Claude Agent SDK provides the same tools, agent loop, and context management that power Claude Code, programmable in Python and TypeScript. An agent built with the SDK completes a task by planning its own steps and calling tools that read files, run commands, or edit code.
The SDK acts as a library that runs the Claude Code binary, providing capabilities such as built-in tools, permissions, sessions, and hooks. The built-in tools allow the agent to read, write, and edit files, run commands, and search the web. The SDK supports connecting external tools and data sources via the Model Context Protocol.
Anthropic does not allow third-party developers to offer claude.ai login or rate limits for products built on the Claude Agent SDK unless previously approved. Developers must use API key authentication methods for the SDK.
According to Claude Agent SDK: The Claude Agent SDK provides the same tools, Agent loop, and context management that power Claude Code, programmable in Python and TypeScript. Its pages list this strength: Provides the same tools, agent loop, and context management as Claude Code. They also state this limitation: Third-party developers cannot offer claude.ai login or rate limits for their products unless previously approved.
β The AI Tool Bible editorial team
Pros
- β Provides the same tools, agent loop, and context management as Claude Code
- β Available in Python and TypeScript
- β Supports hooks to run custom code at key points in the agent lifecycle
- β Allows spawning specialized subagents for focused subtasks
Cons
- β οΈ Third-party developers cannot offer claude.ai login or rate limits for their products unless previously approved.
Facts
- Platforms
- web (not source-checked)
- Pricing
- free (not source-checked)
- GitHub stars
- 2,808 (checked Nov 13, 2026) (not source-checked)
- Last GitHub push
- Aug 27, 2026 (not source-checked)
- First commit
- Sep 2025 (not source-checked)
- Company
- Anthropic, Inc. (not source-checked)
- Parent company
- Anthropic (not source-checked)
- Model used
- Claude Opus / Sonnet (not source-checked)
- Vendor pages
- Changelog
- Hacker News
- 4 stories Β· top 4 points Β· last Oct 28, 2025 Β· checked Nov 9, 2026
Where this tool sits
Use cases
Recent updates from Anthropic, Inc.
Frequently asked
- How much does the Claude Agent SDK cost?
- The tool is listed as free, though no specific pricing details are provided. Developers must use API key authentication, as third-party products cannot offer claude.ai login or rate limits without prior approval.
- What programming languages are supported?
- The SDK is programmable in Python and TypeScript. It acts as a library that runs the Claude Code binary, providing access to the same tools and agent loop used in Claude Code.
- Can I connect external tools or data sources?
- Yes, the SDK supports connecting external tools and data sources via the Model Context Protocol (MCP). It also includes built-in tools for reading, writing, and editing files, running commands, and searching the web.
- What are the main use cases for this SDK?
- Primary use cases include building production AI agents, performing file manipulation and command execution, and connecting external tools via MCP. Agents can plan their own steps and call tools to complete tasks.
- Which AI models power the SDK?
- The SDK utilizes Claude Opus and Claude Sonnet models. It provides the same context management and agent loop capabilities that power Claude Code, allowing for autonomous task planning and execution.