HomeDocsModel support

Docs

Model support

3DLabs routes each pipeline stage to the model that performs best for that task. Model selection is managed by the platform — you don't need to configure individual model endpoints.

Managed Model Access — when enabled in Workspace Settings, the platform provides all AI inference. No personal API keys required.

Image generation

Flux Pro / Flux 2 Profal.ai

Strengths

Highest quality, best prompt adherence, photorealistic detail

Used for

Default for production image generation

Flux Schnellfal.ai

Strengths

Fast generation, good for rapid iteration

Used for

Speed-optimised runs

Ideogram V3fal.ai / Replicate

Strengths

Strong typographic and compositional control

Used for

Scenes requiring text or stylised composition

Gemini ImagenGoogle AI

Strengths

Photorealistic, strong scene understanding

Used for

Documentary-style realism

Video generation

Kling v3 Profal.ai

Strengths

Cinematic motion, high temporal consistency

Used for

Default for production video clips

MiniMax Video-01Replicate

Strengths

Strong motion quality, good character consistency

Used for

Character-driven scenes

Text generation (story & planning)

Gemini 2.5 Flash / 3 FlashGoogle AI

Used for

Story writing, scene planning, TTS script, YouTube metadata

Notes

Primary text model. Fast and cost-effective for all text stages.

Gemini 3.1 ProGoogle AI

Used for

Story writing, scene planning

Notes

Higher quality for complex or long-form topics.

Claude 3.5 SonnetAnthropic via fal.ai

Used for

Story writing, scene planning, TTS script

Notes

Alternative routing option. Strong narrative coherence.

Narration (TTS)

ElevenLabs handles all text-to-speech narration. Each scene gets a narration segment timed to its duration. Narration is generated as part of the Audio Lab stage.