documentation

Developer documentation.

Everything you need to integrate prompt optimization into your stack. SDKs, APIs, and guides for engineering teams.

quick start
01

Getting Started

Set up your first prompt evaluation in under 10 minutes.

coming soon
02

SDK Integration

Integrate with TypeScript, Python, Go, or REST APIs.

View SDKs
03

API Reference

Complete REST and GraphQL API documentation.

coming soon
signature

Clarity in every evaluation.

Understand what changed and why it won.

Prompt Diff
before
You are a helpful assistant. Answer the user question directly.
after
You are a helpful assistant. Answer directly, then add a brief justification in one sentence. Avoid speculation.
delta: +0.22risk: -0.08cost: -12%
Eval Trace
quality
94%
safety
98%
semantic
91%
advanced
87%
Aggregated across 2,847 evaluations · 95% confidence
core concepts

Prompt Registry

Version control for prompts. Create, deploy, and rollback across environments.

  • Version management
  • Environment deployments
  • Template variables

Golden Datasets

Define quality standards with curated input/output examples.

  • Example management
  • Quality benchmarks
  • Import/export

A/B Experiments

Run statistically rigorous experiments on prompt variants.

  • Traffic allocation
  • Statistical significance
  • Winner detection

41D Evaluation

Comprehensive scoring across quality, safety, and semantic dimensions.

  • Multi-model evaluation
  • Custom metrics
  • Scoring thresholds
guides

Multi-Provider Setup

Configure OpenAI, Anthropic, Google, and Cohere providers.

coming soon

AI Optimization

Use AI to generate improved prompt variants automatically.

coming soon

Best Practices

Proven strategies for prompt engineering and optimization.

coming soon

Ready to get started?

Create your first prompt evaluation in minutes.