Hull Gateway

Hull Gateway is the model layer behind Hull. Hull Desktop, the IDE extensions, the Hull CLI and cloud agents use it by default, and you can call it from your own code through an OpenAI-compatible API.

With Hull Gateway you get:

• Many models, one account. Models from Anthropic, OpenAI, Google, xAI, Mistral, MiniMax and others, without separate provider accounts.
• Auto routing. hull-auto/* models pick a model per request. See Auto model.
• Familiar APIs. OpenAI Chat Completions, OpenAI Responses and Anthropic Messages request formats, with streaming.
• Usage tracking. Per-request token and cost records in your Hull dashboard, counted against your plan.
• BYOK. Route requests through your own provider keys. See Bring your own key.
• Organization controls. Pooled usage, per-member spending limits, and model and provider policies on Enterprise.

Base URLCopy Anchor Link

https://api.withhull.com/api/gateway

Every request needs a Hull API key. See Authentication.

ExampleCopy Anchor Link

import { streamText } from "ai"
import { createOpenAI } from "@ai-sdk/openai"

const hull = createOpenAI({
  baseURL: "https://api.withhull.com/api/gateway",
  apiKey: process.env.HULL_API_KEY,
})

const result = streamText({
  model: hull.chat("anthropic/claude-sonnet-5"),
  prompt: "Why is the sky blue?",
})

PlansCopy Anchor Link

Hull Gateway access comes with your Hull plan. Starter and Pro include a monthly AI allowance and Auto routing; Teams shares a pooled allowance across the organization; Enterprise adds managed Hull Gateway with model policies and Enterprise BYOK. See Usage and billing and pricing.

Next stepsCopy Anchor Link

• Quickstart: make your first request.
• Authentication: API keys, organizations and headers.
• Models and providers: model IDs and Auto routing.
• Streaming: server-sent events.
• SDKs and frameworks: integration examples.
• Usage and billing: allowances, limits and controls.
• API reference: endpoints and schemas.