Providers

decide runs on three decision models, through the same commands and templates:

• Jev by TypeSafe, the default, with the jev-latest model:

  export TYPESAFE_API_KEY=...

• Clef by Cloudflare Workers AI, with the clef model, or clef-flash for lower latency. Image templates always use Clef.

  export CLOUDFLARE_AUTH_TOKEN=...
  export CLOUDFLARE_ACCOUNT_ID=...
  export DECIDE_PROVIDER=cloudflare

• GPT-6 Luna by OpenAI, with the gpt-6-luna model, through the Decisions API, which is in beta. Set OPENAI_BASE_URL to send requests somewhere other than https://api.openai.com/v1. Image templates don’t run on it yet.

  export OPENAI_API_KEY=...
  export DECIDE_PROVIDER=openai

Any other service that speaks the Jev API works too, such as one you host yourself. Point decide at it with TYPESAFE_BASE_URL, along with TYPESAFE_API_KEY and a model name:

export TYPESAFE_BASE_URL=https://decisions.example.com
decide run sentiment notes.txt --model my-model

Choose for one run with --provider typesafe|cloudflare|openai and --model NAME, or for every run with DECIDE_PROVIDER and DECIDE_MODEL. --workers sets how many requests run at once (default 4).

Each provider names its own model versions. The answer cache keeps answers for each provider, address, and model name apart.