Descript
An AI audio and video editor with transcript-based editing, transcription, sound cleanup, clips and localization.
OpenAI’s speech-recognition project for multilingual transcription, language identification and supported speech-to-English translation.
Core features
- Transcribe multilingual audio
- Identify spoken languages
- Translate supported speech to English with suitable models
When to use it
Build transcription workflows, turn recordings into text and process multilingual audio.
Plans and practical limits
Code and models are publicly available in the official repository. Local compute, deployment and third-party hosting can incur separate costs.
Dependencies include ffmpeg. The turbo model is not intended for translation tasks. Review important transcripts for recognition errors.
Sources and review
Official source 1 · Checked 19 October 2026. Based on public official sources; signed-in and paid features were not tested.
Relevant Navigation
RAGFlow
An open-source RAG engine combining document parsing, knowledge retrieval and agents to supply traceable context for model applications.
OpenClaw
An open-source AI assistant you can run yourself, connecting tasks, tools and applications through supported chat channels.
HeyGen
An AI avatar-video and translation platform with lip synchronization, voice and visual production features.
LangGraph
A framework for stateful AI agents with customizable workflows, memory, streaming and human-in-the-loop controls.
Adobe Podcast
Adobe’s online AI audio tools for speech enhancement, recording, transcription and document-style audio editing.
LiteLLM
An AI gateway and developer library for multiple model providers, with routing, cost tracking, permissions and usage management.
AnythingLLM
An on-device and self-hostable AI workspace for document knowledge, search and custom agent skills.