Expert-led or self-service

The same platform and the same tools either way. Expert-led always ends with your team running it.

Self-service

Your engineers run the factory on the platform from day one.

  1. Sign up in the web app or from the CLI
  2. Bring a task description and 20 or more examples
  3. Train, evaluate and compare against your current model

Expert-led

Our forward deployed engineers embed in your team and run it with you.

  1. Scope one use case with success criteria agreed up front
  2. We train the first models on your data and read out the results
  3. Your engineers are onboarded and take over on the platform

Which path fits your team

Expert-ledSelf-service
Best forTeams that want results firstTeams that want to own the loop from day one
Who buildsOur engineers, inside your teamYour engineers
We providePlatform and agent, plus scoping, evals, data work and trainingPlatform, agent, onboarding and support
PoC pricingPriced per use casePay as you go, starting with $200 of credit
Model rights in the PoCAgreed in the scopeNon-commercial, hosted with us
Where it endsHandover to your team, in 2 to 3 weeksYou keep building

Get started with existing inputs

Required

Your harness

  • An endpoint to your conversational agent
  • Your system prompt, tools and knowledge base
  • We build mock APIs and test users around it
Helpful

Traces

  • Up to 10k real conversations, no labels needed
  • Requests from a full day to size your deployment
  • Or route n% of traffic through our endpoint
Helpful

Evals

  • The rubric or LLM judge you already trust
  • Hand-labeled scenarios you came across
  • None yet? We agree acceptance criteria together

Choose your way in

Start on the platform today, or talk to us about your use case.