Choose a media API workflow
Map image generation, text to speech, audio transcription, and social video extraction to the UnderSky endpoint, billing model, and recovery path you need.
UnderSky media APIs use the same API key and credit balance as text models. Pick the workflow that matches the input and output you need.
Generate images asynchronously
Use POST /v1/images/generations to create a task. Poll GET /v1/images/generations/:id until it reaches completed, or provide a callback URL.
- Image generation guide explains prompts, reference images, callbacks, and signed URLs.
- Create reference lists request fields.
- Status reference lists task states.
Image tasks reserve credits before work starts. Failed tasks release the reservation, and partial tasks charge for completed images.
Convert text to speech
Send POST /v1/audio/speech with a supported TTS model, input text, voice, and response format. The response is binary audio, so write it to a file or stream it to your media service.
Read the speech reference for models, voices, formats, and character pricing.
Transcribe audio
Send POST /v1/audio/transcriptions with a multipart audio file and a transcription model. Choose a response format that matches your downstream parser.
The audio transcription reference lists accepted fields, model prices, and common errors. The transcription guide shows a complete curl workflow.
Extract social video metadata
Use GET /v1/social-media/video/extract with a share URL or raw share text. The response normalizes platform, title, author, video qualities, audio tracks, and time-limited media URLs.
- Video extraction guide contains supported platforms and examples.
- Video extraction reference contains the response schema.
Download returned media promptly because the URLs expire.
Plan billing and retries
Check pricing before setting budgets. Fixed-price media requests charge per successful unit. Retry only requests that are safe to repeat, and apply the rate-limit recovery rules to 429 responses.
For a first authenticated request, follow the Quick Start. For current availability, check service status.
Choose a model
Choose an UnderSky model by routing behavior, response quality, latency, context needs, and token price, then validate the choice with a small request.
Transcribe audio
Send an audio file to UnderSky's OpenAI-compatible transcription endpoint, choose a model and response format, and handle minimum billing and recovery errors.