Ship AI features
without a backend
Wrap any model into your own endpoint — set a custom prompt, expose only the inputs you choose, then call it from your app. ModelRunner runs the model and stores every output, so there's no server to stand up and no files to host.
Production-ready models, one API
Seedance V1.5 Image to Video
bytedance
Transform static images into dynamic videos with synchronized audio. Supports text-guided animations and start/end frame keying.
LongCat-Video i2v
meituan-longcat
LongCat-Video turns a single still image into minutes-long, smooth 480p, 30fps video with stable style, lighting, and identity — fast, consistent, production-ready animation from one frame.
Real Esrgan Image Upscaler
nightmareai
High-quality image upscaler with optional face enhancement
Seedance 1.0 Pro
bytedance
Seedance 1.0 generates 1080P videos with smooth motion, rich detail, and diverse styles, while the pro version adds multi-shot narrative and advanced instruction following for cinematic results.
Start from a curated collection
Hand-picked sets of models grouped by what you're making — proven line-ups you can browse and run without hunting through the full catalog.
Best Image Models
25 modelsFind the best AI image generator in 2026. Compare top AI image generation and editing models side by side — see real outputs, pricing per image, speed benchmarks, and API integration options.
Image Editing Models
14 modelsEdit images with AI — inpainting, outpainting, style transfer, background removal, and upscaling. Run top image-to-image AI models via API with full parameter control and per-request pricing.
Image to Video Models
7 modelsAnimate still images into high-quality videos. Compare top image-to-video AI models by quality, resolution, audio support, and API pricing.
Audio & Music Models
5 modelsGenerate music, speech, sound effects, and transcriptions with AI. Compare top audio models for text-to-music, text-to-speech, and speech-to-text.
Video Editing & Enhancement Models
3 modelsEnhance, upscale, lipsync, and extend existing videos with AI. Covers video upscaling, audio sync, and video extension.
3D Generation Models
2 modelsGenerate textured 3D meshes (GLB) from text prompts or images, ready for games, AR, and product viewers.
Give your AI assistant the whole catalog
Point Claude, Cursor, or any MCP client at ModelRunner and every model becomes a tool your assistant can call — generate an image, caption a clip, transcribe audio — without leaving the conversation.
{
"mcpServers": {
"modelrunner": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://mcp.modelrunner.ai/sse"]
}
}
}Every file you touch, one click away
Every upload, every input, and every generated output becomes a tracked asset in your library — so a model's result drops straight into the next run with no re-downloading and no re-uploading.
Your first run in three lines
Grab a key, pick a model, send a request. Same shape for every modality, every provider.
$ curl https://api.modelrunner.ai/v1/run \
-H "Authorization: Bearer $MODELRUNNER_KEY" \
-d '{
"model": "black-forest-labs/flux-pro",
"input": { "prompt": "a neon koi, ultra-detailed" }
}'Pay only for what you run
No subscriptions, no idle GPU bills. Per-request pricing — per image, per second of video, or per megapixel — quoted before you run and snapshotted at request time.
Frequently asked
What is ModelRunner?+
ModelRunner is a unified API for running AI models — image, video, audio, 3D, and text — from providers like Google, OpenAI, ByteDance, Stability AI, and Black Forest Labs. One API key reaches every model, and each one uses the same request shape, so switching models is a one-line change.
Which AI models and providers can I run through ModelRunner?+
Hundreds of models across image, video, audio, 3D, and text generation — from labs including Google, OpenAI, ByteDance, Stability AI, and Black Forest Labs, with new ones added regularly. You reach all of them through one API key, with no separate accounts or SDKs per provider.
How do I connect ModelRunner to my AI assistant with MCP?+
Point any MCP client — Claude, Cursor, or your own — at the ModelRunner MCP server with your API key. Every model becomes a tool your assistant can call directly, with no extra integration code.
How do I turn an AI model into my own API endpoint (a wrapper)?+
Create a wrapper: package any model with a preset prompt and only the inputs you choose, and ModelRunner gives you a custom endpoint to call from your app — with no backend to build or maintain. Every output is hosted and saved to your asset library automatically.
How much does ModelRunner cost?+
You pay per request at each model’s published rate — per image, per second of video, or per megapixel. Pricing is snapshotted at request time with no subscriptions or hidden fees, so you only ever pay for what you run.
How is ModelRunner different from calling each AI provider directly?+
One API key and one request shape replace a separate account, SDK, and billing relationship for every provider. You can compare or switch models — even across providers — by changing a single parameter, and ModelRunner hosts every output so there’s no server or storage to manage.
Start building today
One key unlocks every model, every modality, and the MCP server. Free to start — pay only when you run.