Proofreading & Translation

Transform raw speech transcripts into polished, publication-ready subtitles and summaries using Large Language Models (LLMs).

AI Processing Modes

1. Proofreading & Grammar Correction

Fixes speech disfluencies, filler words ("um", "ah", "like"), stuttering, and homophone errors.
Adds appropriate punctuation, capitalizations, and paragraph breaks.

2. Subtitle Translation

Translates transcriptions into 50+ languages while preserving exact subtitle timecodes.
Supports bilingual subtitle generation (Original + Target language side-by-side or stacked).

3. Smart Semantic Segmentation

Splits long run-on sentences into natural, readable subtitle lengths (optimal 5-8 seconds per subtitle block).
Avoids awkward breaks in the middle of names or compound phrases.

4. Meeting & Lecture Summaries

Extracts key takeaways, action items, meeting minutes, and timeline summaries.

Supported LLM Providers

DuRT supports standard OpenAI-compatible API schemas:

OpenAI (GPT-4o, GPT-4o-mini)
Anthropic (Claude 3.5 Sonnet)
DeepSeek (DeepSeek-V3, DeepSeek-R1)
Ollama / Local LLMs (Llama 3.3, Qwen 2.5)
Alibaba Cloud (Qwen-Max, Qwen-Plus)
Zhipu / Kimi / Moonshot / Baichuan