AI-AUGMENTED WORKFLOW AUTOMATION

One Tool. Complete Workflow.

Norse-Powered C++ Performance Engines

Each module is built for maximum performance and designed to handle one part of your video production pipeline perfectly. Local-first. Privacy-respecting.

The Norse Engine Architecture

Loki Studio is built on a modular C++ architecture where each component excels at a specific task. Together, they form the most efficient video-to-YouTube pipeline available.

Module Norse Name Function Performance
Transcription Muninn AI speech-to-text engine 2-22x faster (GPU)
Frame Extraction Huginn Intelligent frame capture Native FFmpeg
Thumbnails Freyja Image compositing engine AVX2 SIMD
Video Metadata Mimir Timeline editing & metadata <5ms reads
Waveforms Heimdall Audio waveform generation SIMD optimized
Audio Analysis Bragi LUFS audio normalization ITU-R BS.1770-4
Clip Extraction Gungnir Stream copy extraction 50-100x faster
Concatenation Bifrost Video segment joining Lossless merge
100% FREE - Local Processing

Muninn - AI Transcription Engine

Convert speech to text using state-of-the-art Whisper AI running entirely on your computer. No API costs, no usage limits, no data leaving your machine.

  • GPU-accelerated with CUDA - 2-22x faster than CPU
  • 99+ languages with auto-detection
  • Real-time translation to English
  • Multi-track audio support - Separate mic + game audio
  • Word-level timestamps for karaoke-style captions
  • Speaker diarization - Identify who's speaking
  • Automatic SRT generation
Muninn Transcription Engine

Available Whisper Models

large-v3-turbo - Recommended
large-v3 - Maximum accuracy
medium / small / base / tiny

AI-Generated Metadata That Actually Works

Generate SEO-optimized titles, descriptions, tags, and chapters from your transcript. Choose from multiple AI personalities to match your brand voice.

AI Metadata Creator

Smart Generation

  • YouTube-optimized limits (100 char titles, 5000 char descriptions)
  • AI-detected key moments and highlights
  • Automatic chapter generation
  • "Make Better" one-click enhancement
  • Edit transcript before generation

LLM Providers

  • Ollama (FREE) - llama3, mistral, qwen
  • LM Studio (FREE) - Any compatible model
  • OpenAI - GPT-4o, GPT-4, GPT-3.5-turbo
  • Anthropic - Claude 3 Opus, Sonnet
  • Custom Endpoint - Any OpenAI-compatible API

15+ AI Personalities

  • Professional - Formal, authoritative
  • Energetic - Exciting, high-energy
  • Mysterious - Intriguing, suspenseful
  • Casual - Friendly, conversational
  • Educational - Informative, teaching-focused
  • Custom - Create your own with JSON
Freyja Thumbnail Creator

Thumbnail Features

AI Frame Selection - Best frames automatically
AI-Generated Text - Personality-driven styling
Game Logo Integration - Brand your thumbnails
Live Preview - Real-time editing

Freyja - Thumbnail Creator

Generate professional thumbnails with AI-assisted frame selection and personality-driven text styling. No Photoshop required.

  • AI-powered frame selection from your video
  • Personality-based styling (fonts, colors, effects)
  • Auto-shrink text to fit
  • WebP and PNG export
  • AVX2 SIMD-optimized compositing

Mimir - Timeline Editor

A Premiere-style timeline editor built for speed. Cut, trim, mix audio, and add captions without the complexity of professional NLEs.

Mimir Timeline Editor

Core Editing Features

  • GStreamer hardware-accelerated playback
  • Multi-track audio waveform visualization
  • Frame-accurate trimming with keyboard shortcuts
  • LUFS audio normalization (ITU-R BS.1770-4)
  • EDL import from Premiere Pro and DaVinci Resolve
  • AutoCut - AI-powered silence removal

Live Caption System

Three Caption Styles

One Word Pop, Full Segment, Sliding Window

Karaoke Highlighting

Word-by-word highlighting as audio plays

Position Options

Bottom, Top, Center, or corner placement

Burn-In or Export

Burn captions into video or export SRT

YouTube Batch Upload

Upload Features

Multi-Channel Support - Multiple accounts
Auto-Import Metadata - Automatic attachment
Queue Management - Reorder and schedule
Resumable Uploads - Network resilient

Batch YouTube Upload

Upload multiple videos with auto-imported metadata, thumbnails, and scheduling. Set it and forget it.

  • Privacy settings (Public, Unlisted, Private)
  • Publish scheduling for optimal release times
  • Playlist management and auto-assignment
  • Progress tracking with detailed status
  • OAuth 2.0 secure authentication

Two Workflows, One Tool

Loki Studio adapts to your editing workflow - whether you're a simple content creator or a professional using advanced tools.

Simple Creators (One-Stop Shop)

If you only need basic cuts and audio adjustments, Loki Studio handles everything:

  1. Record - Capture your gameplay or content
  2. Import - Load video into Loki Studio
  3. Edit - Quick cuts and audio adjustments
  4. Transcribe - AI-powered speech-to-text (FREE)
  5. Generate - AI creates title, description, tags, chapters
  6. Thumbnail - AI-assisted thumbnail creation
  7. Upload - Batch upload to YouTube with scheduling
No other tools needed.

Advanced Creators (Extended Workflow)

Use Premiere or DaVinci for effects, Loki Studio for automation:

  1. Edit in Premiere/DaVinci - Add effects, transitions, color grading
  2. Export EDL - Save your edit timeline
  3. Import EDL - Loki matches your Premiere cuts
  4. Transcribe - Perfect timestamp alignment with your final edit
  5. Generate - AI metadata, thumbnails, captions
  6. Upload - Streamlined YouTube delivery
Why not use Premiere's built-in captions?

They're tedious to work with. Loki Studio provides automated caption sync that just works.

System Requirements

Minimum Requirements

  • OS: Windows 10/11 (64-bit)
  • RAM: 8 GB
  • Storage: 10 GB free disk space
  • CPU: Intel i5 / AMD Ryzen 5
  • GPU: Integrated graphics

Recommended

  • OS: Windows 11 (64-bit)
  • RAM: 16+ GB
  • Storage: 50 GB free disk space
  • CPU: Intel i7/i9 / AMD Ryzen 7/9
  • GPU: NVIDIA RTX series with CUDA
2-22x faster transcription with NVIDIA GPU

For Local LLM (Ollama/LM Studio)

  • Additional RAM: 8-16 GB
  • Storage: SSD (faster model loading)
  • GPU: 8+ GB VRAM (optional, for GPU inference)
100% FREE - No API costs with local models

Self-Contained Installation

  • No external dependencies - Everything bundled
  • Bundled GStreamer - Video playback included
  • Bundled CUDA/cuDNN - GPU acceleration included
  • Bundled Whisper models - Ready to transcribe out of the box
  • Registry-free - Portable-friendly design

Ready to Transform Your Workflow?

From raw footage to YouTube in minutes, not hours. Download Loki Studio and start creating faster.

Download Free View Pricing
☕ Buy me a coffee