Exa Contents

Exa reads a URL through its /contents endpoint and returns the page as clean markdown, or as HTML when you ask for it. It shares the @effect-uai/exa package with Exa search, so the same key powers both.

Install

pnpm add @effect-uai/core @effect-uai/exa effect

Wire it up

import { Config, Effect, Layer } from "effect"
import { FetchHttpClient } from "effect/unstable/http"
import { layer as exaLayer } from "@effect-uai/exa/ExaContents"

const provider = Layer.unwrap(
  Effect.gen(function* () {
    const apiKey = yield* Config.redacted("EXA_API_KEY")
    return exaLayer({ apiKey })
  }),
)

const mainLayer = provider.pipe(Layer.provide(FetchHttpClient.layer))

exaLayer registers two service tags from one implementation:

• ExaContents: the typed tag. Yield this for the knobs below.
• WebRead: the generic tag for provider-portable code.

Calling it

import { ExaContents } from "@effect-uai/exa/ExaContents"

const program = Effect.gen(function* () {
  const reader = yield* ExaContents
  return reader.read({
    url: "https://effect.website/docs",
    format: "markdown",
    maxCharacters: 8000,
  })
})

Every field from CommonReadRequest works here. The typed request adds:

Field	Type	Meaning
livecrawl	never fallback preferred always	Fetch live vs serve from Exa’s cache.
maxCharacters	number	Cap the returned content length.

See also

• Web reading: the cross-provider concept and the portable request / response shapes.
• Exa search: the other capability on this package.