Catalog
4
Image, video, audio, and 3D in one library. Compare price, latency, and example outputs before you run.
Run any model.Ship with one API.
Image, video, audio, and 3D in one catalog. Try prompts in the browser, then call the same model from your app with a single key and transparent credits.
One API Key
Unlimited Possibilities
Platform
One catalog. One key. Every modality.
Stop opening a new account for every generator. Browse image, video, audio, and 3D in one catalog, try a model in Predictions, then call it from your app with a single API key.
API
1
One key for the whole catalog. Same request shape across every model — change the name, keep your code.
Vendors
0
No extra accounts as you add modalities. One invoice, and credits only move when a job completes.
Explore the catalogExplore
The model library.
Video and image models with real provider marks, unit price, and the jobs they run. Try any of them in Predictions.
API
One request shape. Swap the model string.
POST /v1/jobs. The same JSON in cURL, Node, and Python — image, video, audio, or 3D. Change the model name. Keep the rest of your code.
- One API key across image, video, audio, and 3D
- Same JSON — only the model string changes
- Webhooks when the job takes longer than a round trip
model: flux-2-pro · Image
POST /v1/jobs
import ModelMarket from "@modelmarket/sdk";
const mm = new ModelMarket({ apiKey: process.env.MM_API_KEY });
const job = await mm.run({
model: "flux-2-pro",
input: { prompt: "A serene mountain lake at sunrise" },
});
console.log(job.output.url);Workflow
Browse. Test. Ship.
From first prompt to production traffic on one surface.
- Browse
- Test
- Ship
- Scale
- 01
Browse
Compare before you run.
Every card shows an example output, latency, and unit price — so you choose with evidence, not a guess.
- 02
Test
Prove it in Predictions.
Change the prompt, inspect the output, and iterate in the browser. No SDK, no GPU, no deploy.
- 03
Ship
Promote the same model.
A Predictions run is the API call. Swap the model string when you want a different generator — keep the rest of your code.
- 04
Scale
Pay for finished jobs.
Credits move when work completes. When you outgrow the public catalog, host a private checkpoint on the same API.
Hosting
Bring your model. Keep the same API.
Upload a fine-tuned or private checkpoint. You get an endpoint, billing, webhooks, and Predictions — without standing up GPUs.
Live endpoint
Privateacme/brand-voice-v2
Same request shape as every catalog model.
POST /v1/jobs · model=acme/brand-voice-v2- Runs
- 12,480
- Earned
- $186.40
- p50
- 1.2s
Same API
{
"model": "acme/brand-voice-v2",
"input": { "prompt": "Brand intro" }
}Recent runs
- job_8f2a2s ago0.04 cr
- job_3c911m ago0.04 cr
- job_17be6m ago0.04 cr
Upload
Your weights. Our GPUs.
Bring a fine-tuned or proprietary checkpoint. We run inference, scale it, and keep it up from the first request.
Access
Private to your team — or live in the catalog.
Keep the model internal, or publish it so others can run it on credits. You decide who reaches the endpoint.
API
The same request shape as every other model.
Your hosted model uses the same key and payload as the public catalog. Change the name, keep the integration.
Included
Billing, Predictions, and webhooks on day one.
Usage metering, invoices, a browser UI, and completion hooks — ready the moment the endpoint goes live.
Pricing
Pay for completed work.
Prepaid credits. Unit price on every model card. Nothing moves until the job finishes — then you see it on one invoice.
Pay as you go
Credits
no monthly minimum
Top up and run any catalog model. Pricing is shown per unit before every job.
- Predictions
- One API key
- Per-model unit pricing
- Usage dashboard
Production
Usage
scale with traffic
Webhooks, multiple keys, rate limits, and invoices for teams shipping to users.
- Webhooks & polling
- Per-key rate limits
- Project-level spend
- Priority queue
Host
Custom
your model, our GPUs
Upload a fine-tuned or private model. We run compute, billing, and the Predictions UI.
- Private endpoints
- Catalog publish option
- Included billing
- Managed uptime
FAQ
Answers before you integrate.
Pricing, hosting, keys, and how the API stays consistent as the catalog grows.
Model Market is an AI model marketplace and API platform for discovering, testing, integrating, and hosting generative AI models. You get one catalog, in-browser Predictions, and a unified REST API across image, video, audio, 3D, and future modalities.