Familiar
Familiar provides an automated AI dubbing platform that translates video into 30 output languages while preserving the original actor's voice, tone, emotion, and lip-sync. The platform's proprietary diffusion omni-model handles voice, lip movement, and scene audio in a single pass, even with background noise, multiple faces, or hands and microphones in front of speakers. One API request carries a source video of up to 10 GB and 24 hours into up to 29 target languages, with each language rendered independently upon reviewer approval.
- Artificial Intelligence
- Media & Entertainment
- Creator / Creative Economy
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Traditional dubbing workflows require separate voice actors, studios, and quality control per market, making multi-country launches slow and expensive. A 90-minute feature into ten markets costs $63,000 to $135,000 using human dubbing, and the slowest market's schedule delays the entire release. Audiences lose interest when poor voice dubbing is layered on a face that doesn't match, disconnecting viewers from the performance.
Solution
Familiar provides an automated AI dubbing platform that translates video into 30 output languages while preserving the original actor's voice, tone, emotion, and lip-sync. The proprietary diffusion omni-model handles voice, lip movement, and scene audio in a single pass, even with background noise, multiple faces, or hands and microphones in front of speakers. One API request carries a source video of up to 10 GB and 24 hours into up to 29 target languages, with each language rendered independently upon reviewer approval. The platform supports source input in 99 languages and outputs in 30, with each market clearing independently so that approved languages can publish before others are ready. The system includes per-language review steps, Do Not Translate lists, context briefs, and webhook events that track each language's status from submission to completion. Live dubbing for broadcasts is experimental and coming soon, designed to dub streams in real time with each speaker in their own voice.
Target Audience
Primary customers are streaming platforms, film distributors, broadcasters, and content creators who need simultaneous multi-country video launches with consistent voice and lip-sync quality across 30 languages.
Features
- Proprietary diffusion omni-model that renders the speaker's own voice, lip-sync, and scene audio in one model across 30 output languages, handling loud music, crowd noise, multiple faces, and hands or microphones in front of speakers
- One API request supports up to 29 target languages from a single source video up to 10 GB and 24 hours long, accepting input in 99 source languages
- Per-language review workflow with manual approval gates, allowing markets to publish independently as they clear, with each language rendered upon reviewer approval
- Webhook events (dub.output_ready, dub.completed, dub.failed) with signed deliveries and 24-hour retry for each language
- Do Not Translate lists and one-line context briefs persist at the creator scope, surviving re-uploads and master changes
- Automatic source language detection or manual pinning, with translated titles and descriptions included per language via translate_metadata
- Live dubbing support via SRT, RTMP, or WHIP feeds, pushing each language's stream to its own destination (experimental, coming soon)
- Quality tiers (Tier 1, 2, 3) across 30 supported output languages for planning launch priorities
- Free .srt and .vtt subtitle files delivered with every dub, built from the reviewed transcript