AI Media Studio for developers
AI Media Studio has a public REST API, a command-line tool and an MCP server, so you can generate images and videos from your own code or from an AI coding agent. Every call uses the same models and the same workspace credits as the web app. Full reference: developer docs.
Last reviewed .
Does AI Media Studio have an API?
Yes. The REST API lives at https://app.ai-media-studio.com/api/v1 and authenticates with a bearer API key. Workspace owners and admins create keys in the app under workspace settings, API, and choose the scopes each key gets.
| Method | Endpoint | What it does |
|---|---|---|
| POST | /api/v1/images/generate | Generate or edit an image |
| POST | /api/v1/videos/generate | Generate a video from text or an image |
| GET | /api/v1/models | List the image and video models your key can use |
| POST | /api/v1/uploads | Upload a source image for edits or image-to-video |
Key scopes: image:generate, video:generate, media:read, media:write.
curl -X POST https://app.ai-media-studio.com/api/v1/images/generate \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt": "a red panda astronaut, studio lighting"}'Which models can I call?
Image models: Nano Banana 2.1, Nano Banana Pro, GPT Image 2.5, GPT Image 2.6, Muse Image.
Video models: Kling O3 Pro, Kling v3 Pro, Kling 4.0, Seedance 2.0, Grok Imagine Video 1.5.
GET /api/v1/models returns the current list with credit costs, so your code does not need to hardcode it. For a readable overview of what each model costs per image or per second of video, see the AI model price list.
Is there a command-line tool?
Yes. Install the CLI from npm and log in with your API key:
npm install -g @ai-media-studio/cli
aims login --key aims_xxxCan I use AI Media Studio from Claude Code, Cursor or Codex?
Yes. Add the hosted MCP server https://app.ai-media-studio.com/mcp to Claude, ChatGPT, Claude Code, Cursor, Codex or VS Code and sign in; no install or API key needed. In Claude Code:
claude mcp add --transport http --scope user aims https://app.ai-media-studio.com/mcpSetup for every agent, including one-click installs for Cursor and VS Code, is on the AI agent setup page.
Agent skills are also available:
npx skills add liebhardt-io/aims-agentHow is API usage billed?
API calls draw from the same monthly workspace credits as the web app, at the same per-model credit cost. Paid plans start at $19/month.
How do I generate and download an image from the terminal or an AI agent?
One worked session, from a new API key to an image file on disk. Every aims command below comes from the CLI docs and MCP docs.
Create an API key in the app under workspace settings, API. It starts with
aims_. Give it theimage:generatescope (addvideo:generatefor video andmedia:writeto upload your own photos).CLI: install the CLI and log in. The key is stored in
~/.aims/config.json; you can also setAIMS_API_KEY.npm install -g @ai-media-studio/cli aims login --key aims_xxxCheck the key and preview the request without spending credits:
aims whoami --json aims image "a red panda astronaut, studio lighting" --dry-run --jsonGenerate the image.
--print urlprints only the result URL; use--jsoninstead for the full result.aims image "a red panda astronaut, studio lighting" \ --aspect-ratio 16:9 \ --print urlDownload the file from the printed URL with any HTTP client, for example:
curl -o red-panda.png "PASTE_THE_PRINTED_URL"
MCP instead of the CLI: register the server with the claude mcp add command above and start a new Claude Code session, or add this to the MCP config of Cursor, Codex, Claude Desktop or a project .mcp.json:
{
"mcpServers": {
"aims": {
"command": "npx",
"args": ["-y", "@ai-media-studio/mcp"],
"env": { "AIMS_API_KEY": "aims_..." }
}
}
}Then ask in plain language. The agent calls generate_image and generate_video and returns the result URLs:
Use aims to generate a 16:9 image of a red panda
astronaut, then make a 6-second video panning across it.Save the returned URLs the same way as in step 4. For a local photo, run aims upload ./photo.png --print url and pass the returned URL to edit_image or generate_video.
Developer FAQ
Which API key scopes does AI Media Studio have?
AI Media Studio API keys have four scopes: image:generate for POST /api/v1/images/generate, video:generate for POST /api/v1/videos/generate, media:write for POST /api/v1/uploads, and media:read. Workspace owners and admins create keys under workspace settings, API, and pick the scopes for each key.
How do I install the AI Media Studio MCP server?
The quickest way needs no install or key: add the hosted server URL https://app.ai-media-studio.com/mcp to Claude, ChatGPT, Claude Code, Cursor, Codex or VS Code and sign in with your AI Media Studio account. To run the server locally instead, in Claude Code run: claude mcp add aims --scope user -e AIMS_API_KEY=aims_... -- npx -y @ai-media-studio/mcp, then start a new session. For Cursor, Codex, VS Code, Claude Desktop or a project .mcp.json, add a server named aims with the command npx, the arguments -y @ai-media-studio/mcp, and AIMS_API_KEY in its environment. Step-by-step setup for each agent, including one-click installs for Cursor and VS Code: https://www.ai-media-studio.com/agent. The server exposes generate_image, edit_image, generate_video, upload_image, list_models and get_account.
Is there an AI Media Studio command-line tool?
Yes. Install it with npm install -g @ai-media-studio/cli, log in with aims login --key aims_xxx, then run aims image, aims edit, aims video or aims upload. Add --json for machine-readable output, --print url to print only result URLs, and --dry-run to preview a request without spending credits.
What are the AI Media Studio API rate limits?
Every plan has the same generation rate limit: 60 image generation requests and 20 video generation requests per 10 minutes, counted per workspace for API keys. This is a limit on how many requests you can send, not a guarantee of generation speed. A request over the limit gets HTTP 429 with a Retry-After header.
Does the API cost extra?
No. API, CLI and MCP calls draw from the same monthly workspace credits as the web app, at the same per-model credit cost. There is no separate API fee; paid plans start at $19/month.