KrillinAI Released: Open-Source Tool Automating Video Downloads, 100-Language Translation, TTS Dubbing, and Short-Form Conversion
KrillinAI is an open-source pipeline that automates the entire video production workflow, including yt-dlp downloads, Whisper transcription, LLM translation, TT
KrillinAI is an open-source automation tool released in September 2026, designed to maximize efficiency in video content creation and distribution. It automates the entire end-to-end production process—from video downloads using yt-dlp and transcription via Whisper, to LLM-powered multi-language translation, TTS dubbing, short-form re-framing, and cover art generation—all through CLI or manifest-based AI agent workflows.
Image source: KrillinAI / MIKE (@mikenevermiss)
Garnering over 11,000 stars on GitHub and distributed under the GPL v3.0 license, the project supports translation into over 100 languages, facilitating seamless content distribution across global platforms.
Key Features and AI Agent Workflow
KrillinAI modularizes complex video production tasks into a cohesive automation pipeline:
- Comprehensive Video Processing Pipeline: The workflow begins with downloading videos via yt-dlp, proceeds to transcribe audio using Whisper, translates subtitles using LLMs, and completes the production with TTS dubbing.
- Short-Form Optimization: It includes functionality to analyze the key subjects of 16:9 landscape videos and re-frame them into 9:16 vertical short-form formats.
- Manifest-Based Automation: By defining AI agent workflows in the
krillinai_manifest.jsonconfiguration file, repetitive distribution tasks can be easily scripted and automated.
Usage and Caveats
KrillinAI supports both GUI client and CLI environments, providing flexibility for content creators on various platforms, including YouTube, TikTok, Reels, Bilibili, and Douyin.
- System Setup: Users can run Whisper transcription and TTS dubbing models locally or integrate external LLM and TTS API keys as needed.
- Resource Requirements: Hardware resource requirements, particularly local GPU VRAM, vary depending on the resolution and length of the videos being processed. A suitable hardware environment is recommended for high-volume tasks.