Integrations Overview

Pick the right way to connect Atlas Cloud — MCP Server, Skills, CLI, ComfyUI, n8n, or the API directly.

There are several ways to reach Atlas Cloud models. Which one fits depends on where you work.

Choosing an integration

You are…	Use	Why
Coding in an AI-assisted IDE	MCP Server	Your assistant calls models as tools, without leaving the editor
Using an agent that supports skills	Skills	Teaches the agent which model to pick and which parameters matter
Working in a terminal or CI	CLI	Scriptable generation with JSON output and stable exit codes
Building on a node canvas	ComfyUI	Hosted models as drop-in nodes, no local GPU
Automating across services	n8n	Model calls as workflow steps
Writing an application	The API	Full control. OpenAI, Anthropic, and Gemini formats all supported

MCP Server and Skills complement each other rather than compete. The MCP Server executes calls; Skills give the agent the judgment to make good ones. Installing both is the usual setup.

Official integrations

MCP Server

14 tools for image, video, audio, 3D, chat, and account data. Works with Claude Code, Cursor, VS Code, Codex, Gemini CLI, and more.

Skills

Agent Skills for Claude Code, Codex CLI, and Gemini CLI, plus a 25-recipe workflow library.

CLI

The atlas command for generation, model discovery, cost estimation, and account management.

ComfyUI

Drop-in nodes for Seedance, Kling, Wan, Veo, Seedream, FLUX, and more.

n8n

Community nodes for LLM chat and asynchronous media generation.

Using an existing SDK

You do not need an Atlas Cloud SDK for language models. Point the SDK you already use at our base URL:

from openai import OpenAI

client = OpenAI(
    api_key=os.environ["ATLASCLOUD_API_KEY"],
    base_url="https://api.atlascloud.ai/v1",
)

Image, video, audio, and 3D generation use the asynchronous Atlas Cloud endpoints instead — see Predictions.

See LLM API Protocols for what each protocol supports and where behavior differs.

Community projects

Community-built integrations live in awesome-atlas-cloud-integrations, which tracks open-source projects that work with Atlas Cloud. Built something? Open a pull request there.

Community projects are maintained independently and may lag behind the current model catalog. For production workloads, prefer an official integration or the API.