h3 h3 max

h3 max api

h3 max api for key-first video generation

Create a site key, submit an H3 Max request, check queue state, and copy production-shaped examples without exposing the provider credential in browser code.

Text-to-video Image-to-video Server-side proxy D1 key ledger
Diagram showing browser code using a site-issued h3 key through a Worker proxy to H3 Max queue and result APIs.

Step 1

Create a site key

Name Prefix Status Use Revoke
No key loaded yet.

Step 2

Build a generation request

Step 3

Status and result

Ready. Create or select a site key, then submit a request. 
{ "status": "idle" }

Step 4

Copy server examples


          
        

      

    


    

      

        
Plain answer

        
What is h3 max?

      

      

        
h3 max is a developer-facing API gateway for H3 Max video generation. Instead of placing a provider credential in client code, h3 max issues a site key that starts with h3_live_, validates each request at the edge, and then submits approved jobs to H3 Max text-to-video or image-to-video routes from the server. The goal is a small, practical bridge between model documentation and a real integration test: one page for creating a key, sending a prompt, checking queue state, and copying the exact request shape your backend can use.

        
H3 Max is presented by fal as a post-trained variant of MiniMax H3 for higher-quality video generation. MiniMax H3 itself is a multimodal model family described by MiniMax as supporting text, image, video, and audio context with video output that can include synchronized stereo audio. This site keeps those source boundaries visible. It does not claim to be the model creator, it does not expose upstream provider credentials, and it does not treat open-weight MiniMax H3 references as the same thing as the hosted H3 Max API routes. For teams evaluating the h3 max api, the useful first step is to validate the request contract, queue behavior, resolution and duration choices, safety settings, and result payload before building billing, teams, or automation around it.

      

    


    

      

        
API surface

        
Routes included in the homepage workbench

        
The browser only sees site-issued keys. The Worker hashes those keys, stores the key ledger in D1, and sends provider requests from the server when generation is enabled.

      

      

        POST /api/h3-max/keys
        GET /api/h3-max/keys
        PATCH /api/h3-max/keys/:id
        POST /api/h3-max/keys/:id/revoke
        POST /v1/h3-max/generations
        GET /v1/h3-max/generations/:id
      

    


    

      

        
Current request facts

        
H3 Max parameters to verify first

      

      

        

          
Provider model routes

          
minimax/h3-max/text-to-video and minimax/h3-max/image-to-video.

        

        

          
Common fields

          
prompt, duration, resolution, safety_checker, and prompt_expansion_mode.

        

        

          
Resolution

          
Use 480P for quick checks and 768P when visual detail is part of acceptance.

        

        

          
Aspect control

          
Text-to-video accepts ratios such as 16:9, 9:16, and 1:1; image requests should respect the source image shape.

        

      

    


    

      

        
Fit check

        
Use h3 max api when the first proof is integration

      

      

        

          
Good fit

          
Teams that need a fast server-side request path, a clean key boundary, queue status, and a shareable integration example before building a larger video workflow.

        

        

          
Not a fit yet

          
Teams that need full account systems, usage billing, prompt libraries, team roles, or custom fine-tuning in the first release.

        

        

          
Next proof

          
Run a small prompt set, record accepted seconds, check safety outcomes, and confirm result payloads before automating batch jobs.

        

      

    


    

      

        
FAQ

        
h3 max api questions

      

      

        What is the h3 max api?
        
The h3 max api is a key-first gateway for H3 Max text-to-video and image-to-video generation routes. It issues site keys, validates requests, keeps the provider key on the server, and returns queue or result state.

      

      

        Does the browser receive the provider key?
        
No. Browser examples use a site-issued h3_live key. The Worker keeps the provider key server-side and submits validated requests from there.

      

      

        Which H3 Max modes are covered?
        
This workbench covers fal's minimax/h3-max/text-to-video and minimax/h3-max/image-to-video routes, with duration, resolution, prompt expansion, aspect ratio, and image URL fields where those fields apply.

      

      

        Is h3 max the same as MiniMax H3?
        
No. MiniMax H3 is the underlying model family from MiniMax. H3 Max is presented by fal as a post-trained hosted route for H3 Max video generation.

      

    


    

      

        
Source-backed

        
Keep the integration tied to current docs

      

      
Source notes collect the fal API references, fal launch notes, MiniMax H3 materials, and community checkpoints used to shape this homepage.

      Open source notes