goose
an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
open-sourcetool-integration
33.8k
Stars
+1125
Stars/month
10
Releases (6m)
Star Growth
+215 (0.6%)
Overview
goose是一个本地运行的开源AI代理，能够自主执行复杂的软件开发任务。与传统代码助手不同，goose不仅提供建议，还能实际构建完整项目、编写和执行代码、调试错误、编排工作流并与外部API交互。该工具设计灵活，支持任何LLM模型，允许多模型配置以优化性能和成本。goose与MCP服务器无缝集成，同时提供桌面应用和命令行界面两种使用方式。无论是快速原型设计、代码重构还是管理复杂的工程流水线，goose都能适应你的工作流程并精确执行任务，让开发者能够专注于创新而非重复性工作。作为完全开源的解决方案，goose为追求效率的开发者提供了一个强大且可定制的AI助手。
Deep Analysis
Key Differentiator
Unlike Cursor (IDE-embedded) or Cline (VS Code extension), Goose by Block is a standalone autonomous agent that operates at the project level — building, executing, and debugging entire engineering workflows with multi-model configuration and MCP extensibility.
⚡ Capabilities
- • Autonomous on-machine AI agent that builds entire projects, writes/executes code, and debugs failures
- • MCP server integration for extensible tool connectivity
- • Multi-model configuration to optimize performance and cost across different LLM providers
- • Desktop app and CLI interfaces for flexible usage
- • Orchestrates complex engineering workflows including API interactions and pipeline management
- • Custom distribution support for preconfigured providers, extensions, and branding
🔗 Integrations
Any LLM providerMCP serversClaude CodeCodexOpenCode
✓ Best For
- ✓ Developers who want an autonomous coding agent that goes beyond suggestions to execute full engineering workflows
- ✓ Teams building custom AI agent distributions with preconfigured tooling
✗ Not Ideal For
- ✗ Non-technical users wanting a chat interface — use Cherry Studio or Lobe Chat instead
- ✗ RAG or document-centric workflows — use AnythingLLM or RAGFlow instead
Languages
Rust
Deployment
Desktop appCLI (local install)Custom distributions
Pricing Detail
Free: Fully open-source under Apache 2.0
Paid: N/A (bring your own API keys)
⚠ Known Limitations
- ⚠ Requires API keys for LLM providers — no built-in model
- ⚠ Desktop/CLI only — no web UI or server mode for team deployment
- ⚠ Relatively new project by Block — still rapidly evolving
- ⚠ No built-in RAG or knowledge base features
Pros
- + 支持任何LLM模型且可多模型配置，灵活性极高
- + 能够自主完成端到端开发任务，不仅仅是代码建议
- + 开源架构支持自定义扩展和MCP服务器集成
Cons
- - 需要本地安装和配置，对新手用户可能有一定门槛
- - 作为自主代理执行任务时可能需要用户监督和验证结果
Use Cases
- • 从零开始构建完整项目原型，包括代码编写和测试
- • 对现有代码库进行重构和优化改进
- • 管理复杂的工程流水线和自动化开发工作流
Getting Started
1. 通过官方安装指南安装goose到本地环境；2. 配置你偏好的LLM模型和相关API密钥；3. 启动goose（桌面应用或CLI）并描述你的第一个开发任务让它开始执行