Eino
A Go framework for building LLM applications and agents, drawing on ideas from LangChain and Google's Agent Development Kit while following Go conventions, with a graph-based composition layer.
Eino is a framework for building large language model applications in Go, drawing on ideas from LangChain and Google's Agent Development Kit while following Go conventions rather than porting either directly. It defines reusable components, such as a chat model, a tool, a retriever and a chat template, with official implementations for providers including OpenAI, Claude, Gemini, Ark, Ollama and Elasticsearch.
Its Agent Development Kit builds agents with tool use, multi-agent coordination, context management and interrupt-and-resume support for human-in-the-loop steps, including a ready-made agent that runs a tool-calling loop internally, and a DeepAgent mode that breaks a complex task into steps, delegates to sub-agents, runs shell commands or Python code, and searches the web while tracking progress.
A separate composition layer connects components into graphs and workflows that can run standalone or be exposed as callable tools for an agent, so a deterministic data pipeline and an autonomous agent can be combined in the same application.