Open AI models on dedicated GPUs: MasterA AI on ApyHub

Open AI models on dedicated GPUs: MasterA AI on ApyHub

For teams that want Whisper, Flux, and video generation without renting and running GPUs.

01Introduction

MasterA AI runs open-weight AI models on its own GPU hardware and serves them as APIs. Its six APIs cover speech-to-text, chat, text-to-image, and three flavors of video generation. All six are now live on ApyHub, billed in atoms under one subscription, and callable from code or from an AI agent through ApyHub MCP.

MasterA runs its inference on self-hosted GPUs powered by solar energy.

02Why open models on dedicated hardware

Open-weight models like Whisper, Flux, and LTX Video are free to download. Running them is the expensive part. You need a GPU with enough memory, drivers that match the model, a queue for long jobs, and someone to keep it all up. For a team that needs a few hundred transcriptions or a batch of product images a week, that setup rarely pays off.

MasterA handles the hardware side. Its video APIs run on dedicated NVIDIA RTX 5090 cards, and you send a request and collect the result. On ApyHub you get the same models behind one key, one bill, and one set of docs.

Energy is part of the story too. The International Energy Agency's Energy and AI report (2025) estimates that data centers used around 415 TWh of electricity in 2024, and projects that figure to more than double to around 945 TWh by 2030. Providers that run inference on their own renewable supply give teams one way to keep that footprint visible.

03The six MasterA AI APIs

Whisper Speech to Text

Whisper Speech to Text transcribes audio files with OpenAI's Whisper large-v3 model. It detects the language automatically and returns the text, the language code, and the audio duration in one synchronous response. A one-minute recording takes around 8 seconds. Files can be up to 10 MB in WAV, MP3, M4A, and other common formats.

Use it for meeting notes, podcast show notes, voice messages in support tickets, or subtitles.

Chat Completion with Ministral 7B

Chat Completion with Ministral 7B is a small, fast instruction model with an OpenAI-compatible request format. It supports system, user, and assistant roles, replies in 1 to 3 seconds, and has a 6,144-token context window. It costs 1 atom per request.

Use it for short, high-volume tasks: tagging, rewriting, classification, and first-draft replies, where a frontier model is more than the job needs.

Generate Text to Image

Generate Text to Image creates images from a prompt with a choice of two models. Flux.1-Schnell produces images up to 1536 px on either side. SDXL-Turbo is tuned for 512 px and supports a negative prompt to exclude elements. Jobs run asynchronously: you submit, then poll for the result.

Use it for blog and social visuals, product mockups, placeholder art, and ad concepts.

Generate Image to Video with Wan2.2

Generate Image to Video with Wan2.2 animates a still photo into a short MP4, about 5 seconds long, from a motion prompt such as "the subject slowly turns and smiles at the camera". Generation takes 1 to 3 minutes depending on quality and resolution.

Use it to bring product photos, portraits, or illustrations to life for social posts.

Generate Text to Video with LTX Video

Generate Text to Video with LTX Video runs LTX Video 13B and accepts either a text prompt or a source image. Clips run 4 to 6 seconds at 25 fps, at 256 to 1024 px, and a text-to-video job finishes in around 30 seconds. You control width, height, frame count, and inference steps.

Use it when you need fine control over the output format.

Generate Text to Video with LTX-2

Generate Text to Video with LTX-2 runs the larger LTX-2 19B distilled model. You choose a length of 3, 5, or 10 seconds, a quality level (fast, balanced, or best), and SD (480p) or HD (720p). A 5-second clip takes around 30 to 60 seconds.

Use it for short cinematic clips, B-roll, and video ad drafts.

Browse all MasterA AI APIs on ApyHub

04Pricing at a glance

Status checks on the video APIs cost 1 atom each. Image status checks are free.

05For developers: transcribe a file in one call

Whisper Speech to Text is synchronous, so a single request returns the transcript:

bash

curl -X POST "https://api.eu.apyhub.com/mastera-ai/whisper-speech-to-text/v1/whisper/transcribe" \
  -H "apy-token: $APY_TOKEN" \
  -F "file=@/path/to/meeting.m4a"

json

The image and video APIs follow a submit-and-poll pattern. Here is a text-to-image job with Flux.1-Schnell:

curl -X POST "https://api.eu.apyhub.com/mastera-ai/image-generation/v1/flux/generate" \
  -H "apy-token: $APY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Mountain lake at sunset, wide shot", "width": 1024, "height": 768, "steps": 4}'

{
  "status": "processing",
  "message": "Generation started",
  "task_id": "a1b2c3d4e5f6"
}

Poll the task, at most once per second, until it completes:

curl -X GET "https://api.eu.apyhub.com/mastera-ai/image-generation/v1/tasks/a1b2c3d4e5f6" \
  -H "apy-token: $APY_TOKEN"

{
  "type": "flux_t2i",
  "status": "completed",
  "message": "Task completed successfully",
  "task_id": "a1b2c3d4e5f6",
  "output_url": "https://example.com/generated-image.png"
}

06For business users: try it in Claude or ChatGPT

Every MasterA AI endpoint is accessible through ApyHub MCP, so AI agents can discover, evaluate, and call these APIs directly without a hand-written wrapper or tool definition. ApyHub MCP runs in dedicated EU and US regions.

In ChatGPT, add ApyHub EU from the connector directory. In Claude, add ApyHub as a custom connector. Then try prompts like these:

• "Create a 1024 by 768 image of a minimalist desk setup with a plant and a laptop, for a blog header."
• "Make a 5-second video of waves rolling onto a beach at sunset, in HD."
• "Transcribe this voice memo and turn it into a bullet-point summary."

Explore the ApyHub catalog

07Conclusion

Open models give you strong transcription, image, and video generation without a license fee, and the GPU is the hard part. MasterA AI runs those models on dedicated hardware and exposes them as six APIs with fixed per-request pricing. Through ApyHub they sit next to the rest of your stack under one key, and your AI assistant can call them from a chat window.

For another provider spotlight, read how FlowDocs converts PDFs to Excel with rule-based table detection. For more on connecting tools to your assistant, see how to extend ChatGPT's capabilities with plugins.

Start building with MasterA AI on ApyHub

08FAQ

What is MasterA AI? MasterA AI is an API provider that runs open-weight AI models on its own GPU hardware. Its APIs on ApyHub cover speech-to-text, chat, text-to-image, image-to-video, and text-to-video.

Which models does MasterA AI use? Whisper large-v3 for transcription, Ministral 7B Instruct for chat, Flux.1-Schnell and SDXL-Turbo for images, Wan2.2-TI2V-5B for image-to-video, and LTX Video 13B and LTX-2 19B for text-to-video.

How accurate is the Whisper speech-to-text API? It uses Whisper large-v3, the largest Whisper model, and detects the spoken language automatically. Accuracy depends on audio quality, so clear recordings with little background noise give the best results.

How long does video generation take? LTX Video text-to-video takes around 30 seconds. LTX-2 takes 30 to 60 seconds for a 5-second clip. Wan2.2 image-to-video takes 1 to 3 minutes.

Is the Ministral 7B API compatible with the OpenAI format? Yes. It accepts OpenAI-style messages with system, user, and assistant roles and returns a chat.completion object with token usage.

How much does it cost to generate an image? 120 atoms with Flux.1-Schnell and 100 atoms with SDXL-Turbo. Checking the job status is free.

How do I test the MasterA AI APIs? Use the Try it panel on each service page, call the endpoints with curl as shown above, or save the requests in Voiden, an offline, Git-native API client that stores each request as a Markdown file you can version with your code.

Can AI agents use MasterA AI APIs? Yes. Every MasterA AI endpoint is available through ApyHub MCP, so agents in Claude, ChatGPT, and other MCP clients can find and call them.

09About ApyHub

ApyHub is a curated API catalog and trusted operational layer for developers and AI agents. One subscription, priced in atoms, covers the full catalog of over 1,500 endpoints and growing. Every endpoint ships with machine-readable certification aligned with GDPR, SOC 2, and ISO 27001, and is MCP-ready by default. ApyHub is headquartered in Amsterdam, with offices in the Netherlands, Greece, and India, and serves 65,000+ developer workspaces every month. Start on the free tier with no card required.

Building an API of your own? Become an ApyHub provider.