Agenta
Open-source LLMOps: prompt management, evals, and observability.
About
Agenta is an open-source LLMOps platform that bundles prompt management, evaluation, and observability in one place. Prompts get Git-like versioning with variants, branches, and commits; subject-matter experts can edit in the UI without touching code; and the eval side covers LLM-as-judge, custom code evaluators, and human annotation.
Everything is MIT licensed with no feature gating, which is rare in this category. The UI-first design means some workflows feel heavier for teams that want everything in code.
What it can do
- Git-like prompt versioning and environments
- Side-by-side playground testing
- LLM-as-judge and custom evaluators
- Human annotation workflows
- OpenTelemetry-based tracing
- Python SDK and REST API
Setup
- Sign up at cloud.agenta.ai or self-host with Docker.
- Create an app and experiment in the playground.
- Install the SDK: pip install agenta.
- Instrument with ag.init() and fetch versioned prompts at runtime.
Prerequisites
Python 3.8+ for the SDK, or just a browser for the cloud version.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: MIT licensed, no feature gating; Git-like prompt versioning; UI-first; heavier for code-only teams. License: MIT.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 10.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · MIT licensed, no feature gating
- · Git-like prompt versioning
- · UI-first; heavier for code-only teams
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.