Bittensor AI video.Made for
your ideas.
Make films with leading AI models in one creative studio. Behind the scenes, independent miners compete to serve supported requests through EveryFrame’s Bittensor subnet.
What is Bittensor
video generation?
Bittensor connects specialized networks, called subnets, with participants who provide useful services. EveryFrame runs subnet 117 for serving AI media requests. You use the Studio to choose a model, describe a scene and create a video; participating miners supply access to supported models through their provider accounts.
The video models run at their API providers. EveryFrame coordinates eligible requests, and miners handle them from verified cloud workers. This lets independent operators compete on service and pricing while you keep one creative workflow.
You do not need to operate a miner or understand subnet accounting to create a film. Start in the Studio, or integrate generation into your product through the EveryFrame API.
Your idea, your model
Describe the subject, action, camera and setting. Choose a model and its supported reference images, duration and resolution. Review the price before generating.
Miners compete for work
For enabled subnet routes, the coordinator selects an eligible worker with matching capabilities and a current offer. Available capacity and provider access determine which requests can be served.
Build on your result
Review the clip in your workspace. Create another take, bring clips into your timeline, or use Director tools to continue shaping the scene.
For the technical architecture and verification boundaries, read how the subnet works. See network activity on the SN117 dashboard.
Real prompts. Finished films.
Explore different looks made in EveryFrame, from underwater cinematography to stop-motion storytelling. Open any film to watch the output and read its original prompt. These are creative examples from different requests, not a same-prompt model benchmark.
An ocean in motion
Two giant manta rays glide in slow graceful loops over a coral reef, sunbeams dancing through the blue water as schools of silver fish scatter around them.…
Watch the film & read the prompt ↗A fox baker’s morning
Claymation short film about a fox baker. Shot 1: before sunrise the fox in a striped apron kneads dough in a cosy stone cottage kitchen, flour puffing into the…
Watch the film & read the prompt ↗Paper becomes a dragon
A sheet of golden paper on a wooden desk folds itself into an origami dragon, which stretches its wings, breathes a small puff of paper confetti, and flies off…
Watch the film & read the prompt ↗A product reveal
A faceted glass perfume bottle rises slowly out of still black water, a crown of droplets frozen in the air around it, gold rim light tracing its edges.…
Watch the film & read the prompt ↗One balance.
More model choices.
Studio uses credits, with top-ups starting at $5 for 2,500 credits. Your quote depends on the model, length, resolution and reference inputs you choose.
| Model | Resolution | Length | Catalog price |
|---|---|---|---|
| MiniMax H3 Max Turbo | 768p | 5s | 100 credits |
| MiniMax H3 Max | 768p | 5s | 200 credits |
| Kling 4.0 | 720p | 5s | 393 credits |
Checked November 11, 2026. These are catalog examples, not the historical cost of the films above. Prices and availability can change; the quote in Studio is the amount to review for your request. View plans and credits ↗
How miner discounts fit in
Miners publish offers for supported work. For requests using a discounted subnet route, the selected offer can lower the customer quote. The result depends on the route and available offers, so there is no single discount percentage for every model. Customer prices, provider costs and miner rewards are separate figures.
Developers can review dollar-denominated API prices and obtain a request quote in the API model catalog. API funding is separate from Studio credits.
Choose the model for the shot.
An opening-image animation needs different controls from a text-only scene. Check reference modes, supported durations, audio and resolution before choosing. Keep the prompt and settings comparable when testing alternatives.
Plan a scene
Use the Studio and Director to turn an idea into shots, create takes and work toward a finished sequence.
Explore Director ↗Compare model options
Explore the Seedance, Kling, Veo and MiniMax guides for inputs and pricing examples.
Browse models ↗Join the network
Operators can follow the mining documentation to configure provider access and inspect supported work. Creators can simply use the Studio.
Read the mining guide ↗Before your first frame.
Is EveryFrame an AI video platform or a Bittensor subnet?
Both. EveryFrame Studio is the creative application. SN117 is the Bittensor network that participating miners use to serve enabled media requests. The API gives developers access through a programmable interface.
Does every model run on a miner’s own GPU?
No. The supported serving workflow uses provider APIs from verified cloud workers. The model runs at its provider; the miner supplies access and executes the assigned request. A catalog listing alone does not mean that every miner can serve that model.
Can I start with a photo instead of a text prompt?
Yes, when the chosen model supports image inputs. Reference controls differ: some accept an opening frame, while others support first and last frames or additional references. Studio exposes the modes supported by the model you select.
How do I compare the quality of different models?
Start with the same prompt and reference image, match duration and resolution where possible, and review motion, subject consistency and prompt adherence. The gallery helps you explore styles; a controlled comparison requires matched inputs.
Where can I check current model availability?
Use the Studio model selector or the live API catalog. Operators should also check their worker’s supported contracts, active offers and readiness in everycli. Read about model availability.