Live Translate

Fully offline, real-time AI translation. Runs entirely on your own GPU. No cloud. No subscription. Buy once, own forever — every future update is free.

Buy once. Own forever.

Live Translate is a single one-time purchase. Every future update and every new feature is free for existing owners. As open AI technology advances, your copy advances with it — no subscription, no recurring fees, no paywalled features.

Early Access — built together with you

Live Translate is in Early Access and receives regular updates. Your feedback directly shapes what we build next — if something does not work or you are missing a feature, tell us in the Steam Community discussions. We read everything and we answer. Early Access pricing rewards early supporters: as new features arrive, the price may increase. Buy today, and every future feature is yours at today's price.

IMPORTANT — PLEASE READ BEFORE BUYING

Live Translate runs large AI models directly on your computer. Performance scales with your hardware — find your card below and see what you can do:

8 GB VRAM — The full translation suite works
Live voice translation, screenshot translation, batch manga/image translation and video-to-subtitles: all run smoothly on the standard model. For image generation, the fast model (Anime-SM) works; the high-quality model (Anime-XL) is slow at this level. The large translation models (LT Pro / LT Maximum) cannot be loaded.

12 GB VRAM — The full Anime Creator experience
Everything above, plus Anime-XL high-quality image generation and Story Changer run comfortably.

16 GB+ VRAM — Large translation models
All features, plus the LT Pro (26B) large translation model unlocks: more accurate, more natural translations.

24 GB VRAM — Maximum
Everything including LT Maximum (27B, 200+ languages) at the highest settings.

VRAM affects not only which features unlock but also translation speed — more VRAM means faster responses. NVIDIA GPUs are fastest; AMD and Intel GPUs are fully supported via Vulkan, but expect 1.5-3× longer image generation times. Below 8 GB, translation still works but is noticeably slower.

And remember: optimization improves with every update — a feature that pushes your hardware today will run faster on less tomorrow. Owning it today is preparing for the future: your copy gets faster and more capable with every update.

Real-Time Audio Translation

Capture system audio (games, movies, online meetings) or your microphone and watch the conversation appear translated on a customizable on-screen overlay. Powered by an on-device AI translation model with offline speech recognition through multiple interchangeable engines. The app auto-picks the best one for your source language.

• Live translation overlay — drag, resize, recolor, set opacity and font
  
• Speaker identification separates voices in multi-speaker dialogue
  
• Precise Mode lets the AI clean up noisy speech-to-text before translating
  
• 140+ target languages, 28 user-interface languages
  
• Smart cache — repeated phrases translate instantly

Screenshot Translation

Press a global hotkey to capture any region of your screen.

• Inpaint mode — the original text is automatically removed and the translation is rendered in the exact same spot, matching the original layout
  
• Legacy mode — original and translation shown side-by-side

Works on game UI, manga, scanned documents, signs, social media — anything on your screen.

Batch Image Translation

Drop an entire folder of images and translate them all at once. Designed for manga chapters, comic collections, scanned documents, and screenshot archives. Input and output folder selection, optional sub-folder scanning, side-by-side preview, per-file progress.

Video to Subtitles (Vid2Srt)

Drop any video file and Live Translate produces a polished SRT in your target language. The pipeline:

• High-accuracy transcription with built-in hallucination filtering
  
• Two-pass AI translation — per-segment first, then a full-script coherence review
  
• Word-level diff highlighting shows exactly what the AI corrected
  
• Reading-speed targeted segmentation (42 characters × 2 lines, broadcast standard)
  
• Translate one transcription into multiple target languages without re-running speech recognition
  
• Hard-burn into the video with GPU acceleration — NVIDIA NVENC, AMD AMF or Intel QuickSync auto-detected, with CPU fallback
  
• Quality report after each run: coverage, average reading speed, gap detection

Anime Creator

A built-in AI image studio with three modes:

• Text to Anime — generate anime illustrations from a text description
  
• Photo to Anime — turn your photos into anime style
  
• Story Changer — rewrite comic or manga pages with a brand-new story while preserving character consistency

Add speech bubbles, thought bubbles, shout balloons and narration boxes. Save as PNG, copy to clipboard, set as wallpaper, or share to the community gallery in one click.

Steam Workshop Gallery

Browse images shared by other players, vote, favorite, subscribe to creators, and download the original prompts so you can recreate or remix them locally. Share your own creations with one click.

Optional Model DLCs

Live Translate ships with everything you need to start translating immediately. For higher-quality image generation and more powerful translation models, additional model packs are available as DLCs through Steam. Each DLC adds a single capability — install only what you need, when you need it. The app auto-detects installed DLCs and unlocks the matching features automatically.

100% Offline and Private

• No cloud APIs and no third-party AI service dependencies
  
• No telemetry, no analytics, no account login, no required internet connection after install
  
• Audio, prompts, images and translations never leave your device
  
• Runs locally on your CPU and GPU

Supported Platforms

Windows 10 and Windows 11 (64-bit). Multi-language interface in 28 languages.