Collection

Audio, video & music

The media side of the catalogue: speech synthesis for voice interfaces, music generation for content, and video generation with camera control — all metered per request or per character on the same OpenLux key.

Create an API key All collections

Abstract orange glass sculpture for the Audio, video & music collection on OpenLux
Every model in this collection runs on the same OpenLux key and endpoint.

Media models turn the API from a text engine into a content pipeline: a short-form video workflow can script with a text model, narrate with TTS, score with a music model and render with a video model, all under one key and one bill. That is the practical case for a gateway over four separate provider accounts.

Pricing here is per unit rather than per token — per 1K characters for speech and per request for music and video — so cost scales with assets produced, not with prompt length. The rates below include the same discounts as the text catalogue.

The picks

3 models from the live catalogue, with the rate each one is metered at.

Doubao Voice 2

DoubaoTTS

Natural, low-latency speech synthesis for voice interfaces — $0.014 per 1K chars.

Model details

Suno V6

SunoMusic

Full songs and instrumentals for content and short-form video — $0.052 per request.

Model details

Kling 3.0

KlingVideo

Text-to-video and image-to-video generation with camera control — $0.196 per request.

Model details

Audio, video & music questions

Which audio and video models are on OpenLux?

Suno V6 for music and songs, Kling 3.0 for video generation from text or images, and Doubao Voice 2 for speech synthesis with multilingual output.

How are media models priced?

Doubao Voice 2 is metered per 1K characters, Suno V6 and Kling 3.0 per request. All rates are in USD and shown per unit, so asset cost is predictable before generation.

Can I generate video from an existing image?

Yes. Kling 3.0 supports both text-to-video and image-to-video, with camera control for directing shots. The model page and docs cover the request shape.

Do media models use a different endpoint from the chat models?

They run on the same OpenLux base URL and the same API key, with the docs covering the current request format for each modality. One key, one invoice, one place to monitor spend.

Put these picks behind one key

OpenLux meters every model in this collection on the same OpenAI-compatible endpoint, with pay-as-you-go billing and free trial credit to start.

Create an API key Browse all models