Eino Compose

This covers the Eino orchestration layer when you need to wire up multi-step pipelines in Go. You get three APIs: Chain for linear flows, Graph for anything with branches or cycles (like ReAct agents), and Workflow for DAGs where you need fine-grained field mapping between different struct types. All three compile down to the same Runnable abstraction with invoke, stream, collect, and transform modes. The framework auto-converts between streaming and non-streaming nodes, which is genuinely convenient. Most real work happens in Graph since branching and state management are table stakes for agents. The reference docs include a complete ReAct example with tool calling and checkpointing, so you're not starting cold.

skills.sh installs: 352
GitHub stars: 835
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add cloudwego/eino-ext --skill eino-compose --agent claude-code

Installs into .claude/skills of the current project.

Files

Orchestration Overview

The github.com/cloudwego/eino/compose package provides three orchestration APIs:

API	Topology	Cycles	Type Alignment
Graph	Directed graph	Yes (Pregel mode) / No (DAG mode)	Whole input/output
Chain	Linear sequence	No	Whole input/output
Workflow	DAG	No	Field-level mapping

*Chain is implemented on top of Graph in Pregel mode but enforces linear topology.

All three compile into Runnable[I, O] which exposes Invoke, Stream, Collect, and Transform.

import "github.com/cloudwego/eino/compose"

Choosing an API

Chain -- sequential pipeline: prompt -> model -> parser. Simplest API.
Graph -- need branching, loops (ReAct agent), or fan-out/fan-in.
Workflow -- need field-level mapping between nodes with different struct types; DAG only.

Graph Quick Reference

g := compose.NewGraph[InputType, OutputType]()

SKILL.md on GitHub

Details

Repository: github.com/cloudwego/eino-ext
Categories: DevOps & CI/CD, Automation & Workflows, Mobile Development
First seen: Jun 3, 2026

More from cloudwego/eino-ext

All 4