Ingram Cloud

The platform for AI developers.

OpenAI, Anthropic, and Gemini through one API. Keys, usage, billing, and logs in one place. Tools, memory, and retrieval built in.

Get started Read the API docs

Use the SDK you know.

Keep the OpenAI request format across providers. Pick a model per request and inspect the run in the console.

Supported model providers

OpenAI
Anthropic
Google Gemini

This server-side example uses a project token and a published agent. The user ID links the request to its usage and state.

Set up your first request

TypeScript · OpenAI SDK

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.cloud.ingram.tech/v1",
  apiKey: process.env.IC_TOKEN,
  defaultHeaders: {
    "IC-Api-Version": "2026-05-01",
    "IC-Agent-Id": process.env.IC_AGENT_ID!,
  },
});

const response = await client.chat.completions.create({
  model: "anthropic.claude-opus-5",
  user: "user_123",
  messages: [{ role: "user", content: "Summarize this report." }],
});

Connect models. Control spend. Debug requests.

Multiple providers. One API.

Use OpenAI, Anthropic, and Gemini through an OpenAI-compatible interface. Change the model on a request without adding another provider SDK.

Explore model access

Managed keys or your own.

Use managed model access with your Ingram Cloud token, or connect your own provider keys. Manage credentials at project or user level.

How keys work

Usage, costs, and logs together.

Inspect model calls, token usage, and costs in the console. Attribute model spend to users and projects, and set budgets that alert or block new runs.

Usage and budgets

Agent infrastructure included

Add tools and state without building the runtime.

Model access is one part of the stack. Ingram Cloud also runs tools, stores context, and records multi-step execution behind the same API.

Tools and credentials

Connect MCP servers and built-in tools. We handle server-side tool calls, OAuth credentials, and approval pauses.

Read the guide

Memory and retrieval

Persist conversations, add user-scoped memory, and search your documents with managed vector stores.

Read the guide

Hosted agent execution

Configure instructions and tools, publish agent versions, and let the platform run the model-and-tool loop.

Read the guide

Deploy through your app, messaging channels, or scheduled runs. See deployment options.

One place to manage your AI costs.

Managed access combines model charges and the platform fee in your Ingram Cloud balance. Bring your own keys and pay your provider directly, with the platform fee billed by Ingram Cloud. Storage is billed separately.

See pricing

Built for your application.

Separate projects, scoped credentials, and usage attribution by user. Manage configuration through the API, console, or Pulumi provider.

Infrastructure as code Trust and data

Connect your first model.

Create a project, get a token, and follow the quickstart. A verified card unlocks €10 in credit for your first requests.

Get started Read the docs