Guide

What is fabel

Describe a story in chat. An agent writes it, stages it in 3D, voices it, scores it — and fabel renders it into a video you can post.

The core loop

1. Chat — describe your story, or start from a template.
2. Live 3D preview — every change the agent makes updates the player immediately. No manual keyframing.
3. Iterate — "make scene 2 at night", "give the robot a deeper voice". The agent patches the story in place; everything that already works stays as it is.
4. Render — hit Render for an MP4 you can download or share.

Templates

Not sure where to start? The dashboard has six starter stories — pick one and it seeds a full project you can immediately reshape in chat, instead of starting from a blank page.

Voices & audio

Every character speaks with a voice from a curated roster. Ask the agent to cast one (cast_voice) and it generates matching dialogue audio. Regenerating a line you haven't changed costs nothing — audio is cached by its exact text, voice, and settings.

Rendering

Renders run at 720p or 1080p depending on your plan and the project's resolution, and are capped at 3 minutes of output. Free-plan renders carry a small watermark. Every render spends credits from your monthly balance — see pricing.

Sharing

Public share links are opt-in per project — nothing is shared by default. Every shared page has a report button so anyone can flag content for review.

Plans & credits

Every render, voice line, and generated score draws from a monthly credit balance. See/pricing for plan details and credit costs.

Bring your own key (BYOK)

Add your own provider keys under Settings → API keys. Creator lets you bring your own LLM key (Anthropic or an OpenAI-compatible provider); Studio adds your own ElevenLabs key for audio too. BYOK zeroes the platform's LLM/audio credit spend — you pay your provider directly. Render credits still apply either way.