compare
DeepEval vs promptfoo
The same facts for both, read from GitHub every night, and the relation a person reviewed.
| Fact | DeepEval | promptfoo |
|---|---|---|
| Language | Python | TypeScript |
| Licence | Apache-2.0 | MIT |
| Stars | 19k | 26k |
| Latest | python-v4.2.4 | code-scan-action-0.2.1 |
| Last push | 2026-10-05 | 2026-10-06 |
| Release cadence | about 4 days between releases | about 10 days between releases |
| Active contributors | 25+ commit authors on the default branch in the last 90 days | 51+ commit authors on the default branch in the last 90 days |
| Flags | none | none |
How they relate
Both replace Braintrust. Alternatives to Braintrust →
PartialDeepEvalPytest-style evaluations with LLM-as-judge metrics, in Python.
PartialpromptfooEvals and red-teaming from a YAML config, run from the CLI or in CI, with a local web viewer.
DeepEval
- python-v4.2.42026-09-22Introducing Jev in DeepEval
- typescript-v0.9.132026-08-24TypeScript 0.9.13pre-release
- python-v4.2.02026-08-24Python 4.2.0
- typescript-v0.9.112026-08-21TypeScript 0.9.11
- python-v4.1.92026-08-21DeepEval - 4.1.9
promptfoo
- code-scan-action-0.2.12026-10-06code-scan-action: 0.2.1
- 0.124.02026-10-06providers: remove hosted ChatKit provider (#11434)
- 0.123.12026-09-18bedrock: refresh auth across HTTP adapters (#10123) (9d3f063)
- 0.123.02026-09-10providers: default GPT-5.6+ to Responses (#10803)
- code-scan-action-0.2.02026-08-28code-scan-action: 0.2.0
DeepEval
unsigned The latest release, python-v4.2.4, carries no signature GitHub could verify.
Loading the security report
promptfoo
✓ signed The latest release, code-scan-action-0.2.1, carries a signature GitHub verified.
Loading the security report