Agenta

Open-source platform for LLM prompt management and evaluation

Agenta is an open-source LLMOps platform for managing prompts, running evaluations, and debugging LLM application traces. It is designed for teams of developers and domain experts building production AI applications, supporting collaboration throughout the LLM development lifecycle. It is self-hostable and provides tools for prompt versioning, A/B testing, and observability.

Install & run

Dockerdocker compose -f docker-compose.gh.yml up -d

pippip install agenta

Source & docs

Repository Docs

At a glance

Company: Agenta
Pricing: freemium
API available: Yes
Self-hostable: Yes
Launched: 2025-11
Last verified: 2026-08-22

Capabilities

prompt-managementevaluationobservabilityself-hostingcollaborationrag

Alternatives

For AI agents: machine-readable markdown version of this page at /tools/agenta.md, or send Accept: text/markdown.