t3-code-android-nightly/.repos/effect-smol/packages/ai/cloudflare

@effect/ai-cloudflare

Cloudflare Workers AI provider for Effect's DecisionModel. Supports text and JSON classification, ordered ratings, and probabilities with Clef and Clef-flash over REST.

Installation

npm install effect @effect/ai-cloudflare

Usage

Set CLOUDFLARE_ACCOUNT_ID and CLOUDFLARE_API_TOKEN, then provide a Cloudflare decision model:

import { CloudflareClient, CloudflareDecisionModel } from "@effect/ai-cloudflare"
import { Effect, Layer, Schema } from "effect"
import { Decision, DecisionModel } from "effect/ai"
import { FetchHttpClient } from "effect/http"

const Triage = Decision.make({
  input: Schema.String,
  decisions: {
    urgent: Decision.probability({
      instructions: "Does this request need immediate attention?"
    })
  }
})

const Clef = CloudflareDecisionModel.layer({ model: "clef-flash" }).pipe(
  Layer.provide(CloudflareClient.layerConfig()),
  Layer.provide(FetchHttpClient.layer)
)

const program = DecisionModel.decide(Triage, {
  input: "Checkout has been failing for every customer for an hour."
}).pipe(Effect.provide(Clef))

Use "clef" for the larger model. CloudflareDecisionModel.model provides a model descriptor with Cloudflare provider metadata. CloudflareClient.layer accepts an explicit account ID and redacted API token; apiUrl and transformClient customize HTTP access.

The client unwraps Cloudflare's REST envelope and maps failures to AiError. It does not retry automatically. Rate-limit errors include Retry-After when provided. The decision model validates answers and rescales small probability drift from four-decimal rounding.

Cloudflare limits each request to 64 questions, classification to 2–255 choices, and ratings to 2–10 levels. See the Clef API documentation for request limits.