Speech, understood. Real-time multilingual transcription and translation API.
Open format for portable AI characters that stay consistent across models
Stage
Not given
Not given
Founded
Not given
Not given
Based in
UA
IN
Pricing model
Paid
Freemium
Pricing
Not given
Free open-source app (GPL-3.0). OmniChar Cloud: 300 free credits to start. Images priced by resolution: 2 credits at 1K, 5 credits at 2K, 20 credits at 4K (without character). With character: 11 credits at 1K, 18 at 2K, 44 at 4K. Videos priced per second: 10 credits at 480p, 12 at 768p, 26 at 2K, 32
Key features
Supports 100 languages for speech recognition
Translates across 86 languages
Fine-tuned accuracy for Ukrainian, Russian and English
Automatic language detection or explicit language setting
REST API for audio transcription
WebSocket API for live streaming
Translation included at no extra cost
Team management with role-based access on Business and Enterprise plans
Runs on GPU for low latency and CPU for affordable deployment
Portable .char file format for characters across multiple models
Local GPU generation with built-in diffusion engine
Cloud generation service with pay-per-generation pricing
Node-based canvas for building generative pipelines
Embedded character licensing and consent verification
Local LoRA training on 16GB GPUs
Continuity scoring to measure character consistency
Integration with FLUX.2, MiniMax H3, Krea 2, LTX-2.5 and hosted API models
What makes it different
Built-in real-time translation with no additional cost
Proprietary Varmir Engine optimized in-house for accuracy and latency
Supports 100 languages with extra fine-tuning for Ukrainian, Russian and English
Simple API integration for developers and teams
Free and open source application under GPL-3.0 license
Varmir
OmniChar