References and Characters
Choose between text-only generation, up to three character voice references, or one image reference for a FanficLab scene.
References are optional. Use them when they clarify a particular cast or atmosphere; skip them when the scene text already provides enough direction.
Text only
Choose None when you want FanficLab to create the performance directly from the fic scene. This is the quickest starting point and works well when you want to hear the pacing and character dynamic before adding more guidance.
Character voice references
Choose Audio to add up to three character roles. For each role:
- Enter the role or character name.
- Upload a clean audio clip for that role.
- Confirm that the name and clip belong together before generating.
Distinct role names help FanficLab keep dialogue readable in scenes with quick exchanges. A reference can guide identity and delivery, but it does not guarantee an exact copy of a voice across every emotion or line.
Use audio with clear speech and little background noise.
Image reference
Choose Image to upload one visual reference. This mode is useful when the location, lighting, character mood, or overall atmosphere matters more than a specific vocal reference.
A reference image guides the performance; it is not inserted into the audio result. Use a clear image that represents the scene.
One reference type per generation
Audio and image references are separate modes. A generation can use:
- up to three character voice references;
- one image reference; or
- no reference.
You cannot combine character audio and an image in the same request.
When a scene has more than three characters
Only three character voice references can be added. For a larger cast, focus the excerpt on the roles that carry the scene, shorten the moment, or generate from text alone. Do not merge unrelated characters under one role name if readers need to distinguish them.
