Provider keys, costs and your data

Supported providers

Studio accepts API keys from the providers below. You do not need an account with every provider: connect the ones your workflow uses. These links open each provider’s official site or API-key page; sign in or create an account there to get your key.

The descriptions below cover how Studio uses each provider. Available models vary by provider and app version.

Text providers

• OpenRouter (opens in a new tab)

  Text and story preparation using models from multiple companies.

• OpenAI (opens in a new tab)

  GPT text models when selected as your text provider. Also used for content checks.

• Anthropic (opens in a new tab)

  Claude text models and the built-in agent when using Anthropic directly.

Image and video providers

The Seedance labels highlight providers with fewer content restrictions. Restrictions still apply and may change; check the provider’s current terms.

• OpenRouter (opens in a new tab)

  Supported image and video models, using the same key as text.

• fal.ai (opens in a new tab)

  Images and video, plus scene audio and character voices in Audio-First mode.

• kie.ai (opens in a new tab)Seedance: fewer content restrictions

  Images and video.

• MiniMax (opens in a new tab)

  Video through the direct MiniMax API.

• WaveSpeed (opens in a new tab)Seedance: fewer content restrictions

  Images and video.

• Replicate (opens in a new tab)

  Images and video. Replicate calls its key an API token.

• Runware (opens in a new tab)

  Images and video.

• Atlas Cloud (opens in a new tab)Seedance: fewer content restrictions

  Images and video.

Connect the providers you intend to use

1. Use the Keys step or Settings → API Keys to enter the keys requested by your chosen workflow.
2. Check the provider’s available models, balance, and account limits. A provider account may need funded credits before generation works.
3. Start with one small test and review its result and cost before increasing your workload.

Treat API keys as passwords. Do not include them in shared screenshots, support messages, project notes, or agent conversations.

Studio and generation are billed separately

Your Studio access does not include generation credits from the providers you connect. Providers bill your usage under their own plans. A desktop agent may also have its own subscription or usage charges.

Studio’s displayed generation costs are estimates. Check your provider dashboard for actual charges. Retrying or generating alternatives can add cost, so review work before starting large batches.

Check each service’s policies

Generation sends prompts and selected content to the providers used for that task. Depending on the workflow, that can include story text, reference images, audio, or video. A connected AI agent may also send material to its own service.

Review the current privacy, retention, and model-training policies of each provider and connected agent service before sharing content. Policies can differ by product, model, account tier, and settings. Fiction.show does not guarantee that third parties exclude your content from training.

Use material you have permission to share and process. If a provider’s policy does not meet your needs, do not send that material to it.