adk-python and adk-go
About adk-python
google/adk-python
An open-source, code-first Python toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.
Supports model-agnostic agent development with built-in integration for Gemini, MCP tools, and OpenAPI specs, plus hierarchical multi-agent orchestration where coordinator agents delegate to specialized sub-agents. Features a modular architecture enabling deployment to Cloud Run or Vertex AI Agent Engine, with a built-in development UI for testing and an evaluation framework for benchmarking agent performance. Includes session management capabilities like rewinding invocations and tool confirmation flows for human-in-the-loop control.
About adk-go
go-a2a/adk-go
An open-source, code-first Go toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.
Provides hierarchical multi-agent workflows (Sequential, Parallel, Loop) with streaming event-driven architecture using Go 1.23+ iterators. Supports multiple LLM providers (Google Gemini, Anthropic Claude) through a unified interface, extensible tools with automatic function calling, and secure code execution across multiple backends. Includes three-tier session state management, vector-based memory retrieval, and video/audio conversation support for compatible models.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work