In the app’s walkthrough: Connect your AI services · Costs and where your book goes
Before you begin
Choose whether your first exercise is inspection only or a paid generation. Looking through an imported sample does not require generating new media.
Work through it
Open Settings → Connections
Select the provider, enter its API key and choose Save to Keychain. Use the provider named by the model you plan to select. Chapterframe keeps these credentials in macOS Keychain.
Choose a video rendering route
Use fal.ai with its provider connection, or connect a compatible GPU render server in Settings → Render Servers. Vast.ai rentals are configured in the rental window. A provider key and a render-server bridge key are different credentials.
Check the book’s Data & Cost settings
Select the book in the library to inspect its settings and any budget caps. Review the Local or Cloud labels before submitting work so you know where source material will be processed.
Open Costs
The session total in the sidebar opens the Costs window. Inspect spending by chapter and book, and use its CSV export when you want a record. Estimates help plan work; your provider’s billing remains the record of charges.
Back up keys separately
Settings → Connections includes Export Encrypted Keys…. Set a password and keep it separately from the file. Import requires that password. The encrypted export does not include the render-server configuration.
Try it in your chapter
Configure the provider for one planned operation. Locate its cost estimate before running it.
Check before moving on
You know which account pays for the next request and where to inspect its result and cost.
If something goes wrong
A request says the key is missing
Check the provider selected in Connections against the actual model. A key for one provider does not authorize another.
A server connects but cannot render
Inspect its bridge credentials, readiness and supported workflow in Render Servers. Saving a server entry does not install its models.
Still stuck? Contact support with the exact message and the operation you were running.