Ship AI media this week, not next quarter

Stop gluing five vendor SDKs together.

Images from 2 credits (about 2 cents). Video up to 1080p. Tagging for catalogues and moderation. Same request shape for all three, results stored and returned as URLs.

15
models, one key
1 cr
cheapest run (~1¢)
6s
fastest turnaround

Try it — no account

Flash · ~6s · on us

One click:

Straight out of the API

No retouching, no cherry-picked upscales. Each one is a single call with the model and credit cost shown.

Browse all models
Product render of matte black headphones with lime rim lighting
pro-image10 credits
Cinematic portrait of an astronaut in an orbital greenhouse
studio-image5 credits
Neon-lit rainy Tokyo alley at night with a lone cyclist
max-image16 credits
Overhead editorial photo of a sourdough loaf on slate
turbo-image2 credits
1 credit
cheapest generation
~6s
fastest render
15
models, one endpoint
100%
auto-refund on failure

Three studios, one balance

Prototype in the browser, then move the exact same model and options to the API.

Credit cost per generation

No per-model contracts or hidden multipliers. Pick a model, spend its credits.

See plans
ModelCredits
Flash
flash-image
2
Turbo
turbo-image
3
Studio
studio-image
4
Edit
edit-image
14
Pro
pro-image
16
Max
max-image
52
Extend
video-extend
200
Motion
motion-video
40
Motion Fast
motion-video-fast
16
Motion Large
motion-video-large
80
Motion Large Fast
motion-video-large-fast
32
Animate
animate-video
20
Story
story-video
74
Image Tagger
image-tags
4
Image Captioning
image-caption
1

Built for the terminal too

One endpoint for every modality. Send a model slug and a prompt, poll the job or pass wait: true and get the result inline.

  • Scoped API keys

    Create and revoke keys per environment. Usage attributed per key.

  • One job model

    Images, video and tagging share the same request and response shape.

  • Stored outputs

    Results are saved to durable storage and returned as signed URLs.

  • Automatic refunds

    If a generation fails, the credits go straight back to your balance.

POST /v1/generations
curl https://generativeaiapi.com/api/public/v1/generations \
  -H "Authorization: Bearer gai_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "studio-image",
    "prompt": "an orbital greenhouse at golden hour",
    "options": { "aspect_ratio": "16:9" },
    "wait": true
  }'

Before you sign up

What happens if a generation fails?

The credits go straight back to your balance automatically. You are never charged for a job that did not produce an output.

Do my credits expire?

Pack credits never expire. Plan credits refresh with each billing period, and you can cancel any time from your dashboard.

Can I use the outputs commercially?

Yes. Everything you generate is yours to use, including commercially. We do not resell or train on your prompts or outputs.

Do I need a card to start?

No. We're onboarding in batches right now: join the waiting list and 50 credits — around 50 quick images or a first video — land in your account when your access opens. No card.

Fifty credits, no card.

Enough for dozens of images or a first video. Upgrade only when you ship.

Make my first call free