# coze-dev/coze-loop

Next-generation AI Agent Optimization Platform: Cozeloop addresses challenges in AI agent development by providing full-lifecycle management capabilities from development, debugging, and evaluation to monitoring.

Repository: https://github.com/coze-dev/coze-loop
Canonical: https://ross.abutalabs.com/products/coze-loop
Language: Go
License: Apache-2.0
License Family: permissive
Topics: agent, ai, agent-evaluation, agent-observability, agentops, eino, evaluation, langchain, llmops, monitoring, observability, open-source, openai, playground, prompt-management, llm-observability, coze
Last push: 2026-08-26T17:35:13+00:00

## Health v2 (maintenance only)
Score: 74/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 66, longevity 31
- inputs: {"age_days": 436, "days_push": 7, "days_rel": 225, "gap_med": 13, "n_releases_24m": 8}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5705, forks 795 (observed 2026-08-28T04:09:28.120751+00:00)

## What it is
Coze Loop is an open-source, self-hostable platform for full-lifecycle AI agent development, offering prompt engineering with a visual playground, automated evaluation, and end-to-end observability of agent execution. It is written in Go and provides a web-based developer platform for debugging, evaluating, and monitoring LLM-powered agents.

## Use cases
- manage and version prompts for LLM apps
- debug and compare prompts across different models in a playground
- run automated evaluations on agent outputs
- trace and monitor agent execution from input to output
- capture exceptions and intermediate results in LLM pipelines
- self-host an LLMOps platform for agent development

## When to choose
- you need prompt versioning, debugging, and comparison tooling
- you want automated multi-dimensional evaluation of prompts and agents
- you need full-trace observability for LLM agent execution
- you want a self-hosted, extensible agentops platform

## When to avoid
- you only need a simple LLM API client or chat UI
- you need a fully managed cloud service with no self-hosting overhead
- your stack requires a platform in a language other than Go

## Facets
- artifact type: service
- maturity: active
- function: monitoring, prompt-engineering, agent-framework, llm-inference, tracing, testing, developer-tools
- domain: large-language-models, developer-tools, monitoring
- platform: self-hosted, go, cross-platform
- tags: llmops, agentops, prompt-management, agent-evaluation, observability, playground, tracing, ai-agents, natural-language-processing, docker, web-server

## Member repositories
- coze-dev/coze-loop (main) score 74

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:28.120751+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:53:33.277771+00:00, confidence not recorded.
  - readme: https://github.com/coze-dev/coze-loop (fetched 2026-08-28T04:09:28.120751+00:00, sha 7cef7bba0d68)
- Data as of 2026-08-30T08:39:29.467469+00:00.