MCP server · Claude · ChatGPT · Codex
Make images and video without leaving Claude or ChatGPT
Connect Gizmo once and your assistant can generate with 15 image models and 27 video models, including Seedance, Veo, Kling, Nano Banana and GPT Image. Images appear in the chat; videos come back as an MP4 link.
Image and video generation need the Starter plan or above. See plans
How it works
Connect once
Add Gizmo as a custom connector and sign in. There is no key to copy: each app gets its own, and you can disconnect it any time.
Ask in plain words
Describe what you want and, if you like, which model. The assistant picks the settings, passes along your reference images and handles the rest.
Get the file
Images come back in the chat with a link. Videos take 1–5 minutes; the assistant checks on them and hands you the MP4.
Connect your app
Every app uses the same server URL:
https://gizmoplatforms.com/api/mcpClaude
Settings → Connectors → Add custom connector. Paste the URL, click Connect, sign in to Gizmo and allow access.
ChatGPT app / Codex
Plugins → MCPs → Add → Connect to a custom MCP. Set Type to Streamable HTTP (not STDIO), paste the URL, save, and sign in to Gizmo when asked.
Claude Code, Cursor and agents
Clients that can send headers can use a Gizmo API key instead of signing in:
claude mcp add --transport http gizmo https://gizmoplatforms.com/api/mcp \
--header "Authorization: Bearer $GIZMO_API_KEY"STDIO clients, the Codex config file and the full tool list are in the MCP server docs.
Things to ask
Name a model or let the assistant choose. A few that work well:
Make a 3:4 poster of a lighthouse at dawn with Ideogram 3.
Turn this photo into an 8-second Veo 3.1 clip with sound.
Compare Nano Banana Pro and GPT Image 2 on the same prompt.
Write a 15-second ad script for my candle shop, then make the opening shot with Seedance 2.5.
Give me four square product shots of this sneaker on different colored backgrounds.
Animate this logo: start from the image, end on a plain white background, 5 seconds.
Every Gizmo model, one connection
The assistant can list the models and their prices before it generates, so you can ask for the cheapest one or the best one.
How billing works
- Generations are billed to your Gizmo account at the provider’s metered price plus your plan fee, the same as the API and the playground.
- Your plan’s weekly included usage is used first, then your credit balance.
- With auto-refill on, a generation that needs more than your balance tops it up from your saved payment method and keeps going.
- Settings → Connected apps shows what each app has spent and made, and Disconnect cuts off its access at once.
Questions
What is MCP?
The Model Context Protocol is the open standard Claude, ChatGPT and other assistants use to call outside tools. Gizmo runs an MCP server, so they can use its image and video tools.
Which plan do I need?
Image and video generation need Starter ($20/month) or above. Starter includes $5 of usage every week; past that, usage draws from your credits.
What can the assistant do with my account?
Only what the tools allow: generate images and video, list your past generations, and use computers you connected to Gizmo with control turned on. It never sees your password or card.
How long do videos take?
Usually 1–5 minutes. The assistant waits about 75 seconds, then checks back until the MP4 link is ready.
How do I disconnect?
Settings → Connected apps → Disconnect. The app loses access right away and has to sign in again to reconnect.
Bring every model into your chat
Create an account, choose Starter or above, and connect Claude or ChatGPT.