One key for image, video, voice and music. No plan.
The media layer for products and agents: 144 models, lip-sync included, on one balance. Every call can be priced before it runs, caps per key and per project stop a runaway job, and your balance never expires.
GPT Image 2 · Nano Banana · Veo 3.1 · Seedance · Kling · ElevenLabs · Suno · Mureka · and 136 more
curl -X POST https://api.fairstack.ai/v1/generations/image \
-H "Authorization: Bearer $FAIRSTACK_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"gpt-image-2-t2i","input":{"prompt":"A mountain lake"}}'
# Response
{
"id": "gen_abc123",
"url": "https://media.fairstack.ai/...",
"model": "gpt-image-2-t2i",
"cost": {
"amount": 0.0636,
"currency": "USD"
}
}One integration. Every modality.
Every new modality means another provider, another dashboard, another billing system, another set of docs.
- × One vendor and key for images
- × Another for video
- × Another for voice
- × Another for music, each with its own bill
- ✓ One API key for all modalities
- ✓ One billing dashboard in real dollars
- ✓ One set of docs, one SDK
- ✓ Switching models = changing one parameter
Three lines to generate anything.
Same SDK, same pattern, every modality. Switch from image to video by changing the method call.
import FairStack from "@fairstack/sdk";
const fs = new FairStack();
const image = await fs.image.generate({
model: "gpt-image-2-t2i",
prompt: "A mountain lake",
});
console.log(image.url);
// cost: $0.0636/image (medium, 1:1)import FairStack from "@fairstack/sdk";
const fs = new FairStack();
const audio = await fs.voice.generate({
model: "elevenlabs-multilingual-v2",
text: "Welcome to the show",
});
console.log(audio.url);
// cost: $0.216/1K charsimport FairStack from "@fairstack/sdk";
const fs = new FairStack();
const video = await fs.video.generate({
model: "kling-3-0-std",
prompt: "A timelapse of clouds",
});
console.log(video.url);
// cost: $0.12/s
Also available: pip install fairstack for Python
APIs for every creative modality.
One API key. 144 models across image, video, voice, and music. Switch models in one line of code.
Image Generation
/v1/generations/image 39 models including Z-Image Turbo, FLUX.2, Seedream 4.5, Nano Banana Pro, GPT Image 2. Text-to-image, image-to-image, inpainting, upscaling.
Base model cost + a flat 20% platform fee, shown before you spend
Video Generation
/v1/generations/video 62 models including Veo 3.1, Kling 3.0, Seedance 2, Wan 2.6, Hailuo 2.3. Text-to-video, image-to-video, video extend, restyle.
Base model cost + a flat 20% platform fee, shown before you spend
Voice & TTS
/v1/generations/voice 10 models including ElevenLabs Multilingual V2, Fish Audio S2.1 Pro, Qwen3-TTS, IndexTTS2. Text-to-speech, speech-to-text, dubbing, emotion control. 168 built-in voices.
Base model cost + a flat 20% platform fee, shown before you spend
Music Generation
/v1/generations/music 19 models including Suno, Mureka V9, Mureka BGM. Text-to-music, lyrics-to-song, instrumentals, extend and edit.
Base model cost + a flat 20% platform fee, shown before you spend
No credits. Just dollars.
Every API response includes the exact cost in USD. Base model cost + a flat 20% platform fee = your price. We publish the math.
- Every response includes cost in USD
- Call POST /v1/estimate, or send confirm: false, before generating
- No credits to decode, no token math
- Flat 20% platform fee on base model cost, published formula
| Model | FairStack |
|---|---|
| GPT Image 2 | $0.0636/image (medium, 1:1) |
| ElevenLabs Multilingual V2 | $0.216/1K chars |
| Veo 3.1 Fast I2V | $0.18/s (720p) |
| Kling 3.0 Standard | $0.12/s |
Not just another aggregator.
Full cost controls, structured errors, and real documentation.
Scoped API Keys
Create keys with granular permissions: restrict by modality, set per-request cost limits, set total spending caps. Revoke instantly.
3-Level Spending Caps
Hard limits at organization, project, and API key level. A request over a cap is refused before it runs, not billed. Safe for autonomous agents.
Webhooks
Register endpoints per org. Get notified on generation.completed, generation.failed, job.completed and job.failed events. Automatic retries with delivery logs.
Cost Simulation
Call POST /v1/estimate before generating. Get the exact dollar price, choose the model, then execute.
Structured Errors
Every error includes a machine-readable code, human message, and request trace ID. No guessing what went wrong.
MCP Server
Agent-native integration for Claude, GPT, custom agents. Your AI can generate images, video, and audio as tool calls.
Ship in minutes.
REST API, webhooks, SDKs, and MCP — everything you need to ship.
What happens to money you don't spend.
Every option here is pay as you go. The difference is what happens to a prepaid balance when a project pauses, and whether one balance covers your app, your API calls and your agents. Unit prices vary by model and are on each model page.
| Who | Prepaid balance expires | How they charge |
|---|---|---|
| FairStack | Never | Flat 20% platform fee on base model cost, per generation |
| OpenRouter | May expire after one year | 5.5% (Standard) or 8% (Business) when you buy |
| Replicate | One year | Per-model prices |
| fal | 365 days | Per-model prices |
| Higgsfield API | One year after funds are added | Per-generation prices, $5 minimum |
| WaveSpeedAI | Never | Per-use prices |
Read on each vendor's own page on 2026-10-31: OpenRouter FAQ · OpenRouter pricing · Replicate prepaid credit · fal terms, 9(a) · Higgsfield API help · WaveSpeedAI pricing docs
Frequently asked questions
What can one FairStack API key do? +
Generate images, video, speech, music and lip-synced talking video, transcribe and dub audio, across 144 models, from one prepaid balance. The same key works over REST, the SDKs and the hosted MCP server.
How do I know what a call costs before I make it? +
Call POST /v1/estimate with the model and settings, or send the generation with confirm set to false. You get the dollar price and nothing is charged. Every price is base model cost plus a flat 20% platform fee.
Can I cap spend per customer, project or agent? +
Yes. Each API key can carry a per-request cap and a total cap, and each project can carry a monthly cap. Requests over a cap are refused before they run.
Is there a plan or a minimum spend? +
No. FairStack has no plans. Add funds with a $1 minimum. Your balance never expires, and the same balance covers the web app, the API and MCP.
Does the API include music? +
Yes. Suno-family and Mureka models generate songs, lyrics and instrumentals on the same key as image and video. See the music API page for prices.
Still have questions? We're here to help.
Ship AI features this week.
144 models. One API key. No subscription. No monthly fee.