For the complete documentation index, see llms.txt. This page is also available as Markdown.

Three experiences (Public · Premium · SDK)

Stratix Public, Stratix Premium, and the Python SDK — what each one is for and how they work together.

LayerLens Stratix is delivered as three distinct customer-facing experiences. They share one catalog, one judge engine, and one trace pipeline — but each one is shaped for a specific kind of user and a specific kind of question.

Side by side

Stratix Public
Stratix Premium
Python SDK

Where

stratix.layerlens.ai — signed out

stratix.layerlens.ai — after you sign in

pip install layerlens --extra-index-url https://sdk.layerlens.ai/package

Auth

none

OAuth (Google, GitHub) + email

API key (X-API-Key)

Audience

researchers, buyers, the curious

builders, operators, admins

builders, automation, CI/CD

What you do

browse, compare, read

build, evaluate, govern

automate, integrate, batch

Catalog visibility

full public catalog

full public + private

full public + private

Private evaluations

no

yes

yes

Build judges

no

yes

yes

Trace upload

no

yes

yes

GEPA judge optimization

no

yes

yes

Org and team management

no

yes

n/a

Cost

free

free + ECU PAYG

free + ECU PAYG

Stratix Public — the showcase

Anonymous browsing. No account required. Built for researchers, AI buyers, and anyone who wants to know "which model is actually best at X right now."

You can:

  • Browse 175+ models with metadata, benchmark scores, and the latest leaderboard

  • Browse 52+ benchmarks with descriptions and per-model scores

  • Browse 2,000+ public evaluations

  • Compare any two models head-to-head across any shared benchmark

  • Browse public evaluation spaces — curated bundles of model + dataset + scoring

  • Search the entire catalog

  • Read the quarterly research reports

You cannot:

  • Run a private evaluation on your data

  • Save or share work to a personal workspace

  • Build judges or scorers

  • See trace data

Stratix Public is the doorway. Most users start here, find a model or a benchmark they care about, and continue into Premium when they're ready to evaluate against their own data.

Stratix Premium — the workspace

Logged-in. Multi-org. Pay-as-you-go on compute. Built for teams shipping AI features.

You can:

  • Run private evaluations on your own data

  • Build LLM judges for subjective dimensions and optimize them with GEPA

  • Upload traces and run trace evaluations against scorers and judges

  • Manage scorers (code graders) at the org level

  • Run agentic evaluations — pre- and post-deployment quality gates for multi-step agents

  • Use BYOK custom models — register your own OpenAI-compatible endpoint

  • Manage organizations, multi-org membership, and team seats

  • Buy and consume ECU credits

  • Browse the Learning Library and follow Paths

The left rail is: Home, Evaluations, Models, Benchmarks, Scorers, Spaces, Agent Evaluation (Traces, Judges), Learning (Library, Paths), Settings, Notifications. ECU balance and Buy Now appear in the header. The in-app Assistant offers context-aware help. The Switch toggle lets you flip between the Public and Premium catalogs without losing context.

Python SDK — the automation surface

pip install layerlens --extra-index-url https://sdk.layerlens.ai/package (v1.3.0). 11 resources, 138 sample programs.

You can:

  • Programmatically create evaluations, judges, scorers, traces

  • Run trace evaluations as part of CI/CD

  • Sync results into your own data warehouse or BI

  • Automate judge GEPA optimization

  • Integrate with frameworks (LangChain, LlamaIndex, Haystack, etc.) via instrumentation samples

The SDK is sync-and-async. Both client.evaluations.create() and await client.evaluations.acreate() work.

How they work together

A typical team uses all three:

  1. A researcher browses the public catalog and picks the three candidate models that look strongest on a relevant benchmark.

  2. A builder logs into Premium, runs a private evaluation on the team's own dataset, and chooses a winner. They build a judge that captures the team's quality bar.

  3. A platform engineer wires the SDK into CI/CD: every PR that changes a prompt re-runs the evaluation. Judges and scorers gate the merge.

  4. An operator ingests live traces and runs trace evaluations on a daily cadence. Quality regressions surface in a dashboard before they surface to customers.

Where to next

Last updated

Was this helpful?