// agents · coze-dev/coze-loop

coze-loop

coze-loop é um projeto de agentes de IA de código aberto de coze-dev (coze-dev/coze-loop no GitHub). Tem 5.765 estrelas no GitHub e 801 forks, é publicado sob a licença Apache-2.0 e foi atualizado pela última vez no GitHub em 8 de outubro de 2026.

A full-lifecycle platform for developing, debugging, evaluating and monitoring AI agents.

★ 5,8 mil estrelas⑂ 801 forksApache-2.0atualizado out 2026

Ver no GitHub ↗

▦ visual: o README deste projeto não inclui nenhuma captura incorporável,
então aqui vai um marcador. O guia abaixo está completo mesmo assim.

// dados do projeto

Dados do projeto

Repositório: coze-dev/coze-loop
Estrelas no GitHub: 5.765
Forks: 801
Licença: Apache-2.0
Última atualização no GitHub: 8 de outubro de 2026
Categoria: Agentes de IA
Tipo de ficha: Guia de instalação completo

// como funciona

Como funciona

Clarify your agent quality problem
Identify what you need to improve — prompt quality, evaluation coverage, or visibility into failed runs.
Deploy Coze Loop
Clone the repo and prepare the Docker Compose deployment under release/deployment.
Configure a model
Edit model_config.yaml with your model provider's API key and endpoint/model ID.
Start and open the UI
Bring the services up and open the web interface to develop prompts, run evaluations and inspect traces.

// recursos

Por que as pessoas gostam tanto

Prompt development playground

Write, debug, optimize and version prompts in a visual playground with real-time interactive testing and side-by-side comparison.

Systematic evaluation

Automated multi-dimensional evaluation of prompts and agent outputs — accuracy, conciseness, compliance and more.

End-to-end observability

Records the full execution process from user input to AI output, including prompt parsing, model invocation and tool execution, capturing intermediate results.

Full lifecycle in one platform

Development, debugging, evaluation and monitoring live together instead of in separate tools.

Flexible deployment

Runs via Docker Compose for a quick local deployment or via a published Helm chart on Kubernetes.

Built on Eino

LLM integration is provided by the CloudWeGo Eino framework team, with a Go backend.

// execute você mesmo

Execute você mesmo

Clone the repository
Clone Coze Loop and enter the project directory.
```
git clone https://github.com/coze-dev/coze-loop.git
cd coze-loop
```
Configure your model
Edit the Docker Compose model config with your provider's API key and model/endpoint ID.
```
# Edit release/deployment/docker-compose/conf/model_config.yaml — set api_key and model
```
Start the service
Bring Coze Loop up with the provided make target and open the UI.
```
make compose-up
# Then open http://localhost:8082
```

// casos de uso

O que você pode criar

Comparing prompt variants in a playground before shipping an agent change
Running regression evaluations over a dataset whenever an agent is updated
Tracing a failed agent run step by step to find where it went wrong
Giving a team one shared place for agent development, evaluation and monitoring

// perguntas frequentes

Perguntas frequentes

Who maintains Coze Loop?

Coze Loop is maintained by coze-dev, the team behind Coze, with the Eino framework providing LLM integration.

Is Coze Loop free?

The open-source version is Apache-2.0 licensed and free to self-host; Coze also offers hosted products separately.

Does it need a paid API key?

Yes — you configure a model provider (the examples use Volcengine Ark / BytePlus ModelArk) with your own API key in model_config.yaml.

How does it compare to Langfuse or LangSmith?

Those focus mainly on tracing and evaluation; Coze Loop covers the full lifecycle including a prompt development playground and version management, in a self-hostable Go platform.

How do I get started?

Os outros projetos de Agentes de IA com mais estrelas no catálogo do LibreAI — a lista realista de alternativas se coze-loop não for o ideal para você.

openclaw★ 391,6 mil
A personal AI that really does things — on any OS, with your data staying yours.
superpowers★ 294,9 mil
An agentic skills framework & software development methodology that works.
ECC★ 275,2 mil
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
hermes-agent★ 252,0 mil
The agent that grows with you — a Python framework for adaptive AI agents.
deepseek-harness★ 245,5 mil
DeepSeek's plugin-based agent harness where everything is a plugin.

// mais em agents

Mais em Agentes de IA

Projetos de Agentes de IA próximos no catálogo, os mais próximos em estrelas de coze-loop: continue explorando dos dois lados dele.

pyspur★ 5,8 mil
A visual playground for agentic workflows: Iterate over your agents 10x faster
ruoyi-ai★ 5,7 mil
Enterprise-grade AI agent framework with multi-provider LLM management, secure knowledge bases and high-precision RAG, visual workflow orchestration, and mul…
klavis★ 5,8 mil
An MCP integration platform that lets AI agents use tools reliably at any scale.
eve★ 5,5 mil
Vercel's open framework for building agents that run anywhere.
agentops★ 5,9 mil
Python SDK for AI agent monitoring, LLM cost tracking, benchmarking, and more. Integrates with most LLMs and agent frameworks including CrewAI, Agno, OpenAI …
TaskingAI★ 5,4 mil
An open-source platform for AI-native application development — assistants, tools, RAG and models in one API.

Baixe o coze-loop

Grátis e aberto: pegue o código-fonte, dê uma estrela no repositório e comece a criar.

Ver no GitHub ↗← Mais ferramentas de Agentes de IA