GEPAzilla

Spec sheet

Purpose

GEPAzilla is an open-source GEPA prompt optimizer that uses datasets and custom metrics to find stronger prompts.Read 10 Oct 2026 · gepazilla.com

Intended users

The maker describes the console as a way for curious builders to experiment with prompt optimization using their own API key.Read 4 Oct 2026 · github.com

Deployment

The README describes the application as a standard Next.js project that can be built and started locally.Read 10 Oct 2026 · github.com

License

The repository states that GEPAzilla is released under the MIT License.Read 10 Oct 2026 · github.com

API key requirement

The project README says an AI Gateway API key is required for GEPA runs, scorer previews, and model discovery.Read 10 Oct 2026 · github.com

Built-in latency metric

GEPA optimizes request latency internally, and the Results tab always includes a latency metric.Read 10 Oct 2026 · gepazilla.com

Data handling

The README says telemetry stays in the browser, telemetry events are not sent to a remote service, and run data is not exfiltrated.Read 10 Oct 2026 · github.com

Data import limit

Dataset tools currently support copying or pasting JSON payloads from the clipboard.Read 4 Oct 2026 · github.com

Dataset format limit

Dataset tools currently support copying or pasting JSON payloads from the clipboard.Read 10 Oct 2026 · github.com

Dataset input

Users can load their own rows or start with the sample dataset.Read 10 Oct 2026 · gepazilla.com

Diagnostics

Results include scorer failure rates, averages, and common notes, alongside candidate prompts and iteration trends.Read 4 Oct 2026 · github.com

Gateway integration

The README specifies a Vercel AI Gateway credential for the required API key.Read 10 Oct 2026 · github.com

Model access

The project requires users to supply an AI Gateway API key for GEPA runs, scorer previews, and model discovery; its README specifies Vercel AI Gateway.Read 4 Oct 2026 · github.com

Scoring

It supports combining deterministic checks with LLM judges and weighting or duplicating scorers.Read 10 Oct 2026 · gepazilla.com

Telemetry

The maker says telemetry stays local and includes tokens, routing, costs, and errors for inspection.Read 10 Oct 2026 · gepazilla.com

Validation

Validation rows are kept out of reflection and used to rank candidate prompts on the Pareto frontier.Read 4 Oct 2026 · gepazilla.com

Validation data

Validation rows are kept untouched during reflection and used to rank candidates on the Pareto frontier.Read 10 Oct 2026 · gepazilla.com

Where Specifiction places it

• AI Prompt Generators#20 of 287.3