> montajo --intro · local-first ai video studio

Ship Viral Shorts & Edit in 1.2 Seconds

// Turn raw topics, campaign briefs, or long footage into captivating 9:16 vertical videos and complete channel packages. Narrated by Kokoro’s 28 natural voices, synced to Remotion motion graphics, rendered with 3D Blender stagecraft, and updated with zero untouched re-encoding.

[ ▶ launch studio ] [ explore the engine → ]

[ 03 ] // workspaces

[ 1–2s ] // delta updates

[ 28 ] // kokoro voices

[ 100% ] // local whisper

⚡

Hook Score 98 00:00.0 → 00:04.2

◎

QA Verdict: PASS Face + Subs + #ad

✓

1.2s Delta Ready 214.6K Shorts Views

AI FOLLOW · PRO

COURAGE IS THE INPUT

00:00:16:57

montajo-cli — local-pipeline · zsh ZERO RE-ENCODE

╔══════════════════════════════════════════════════════════════════════════════════╗
║  ███╗   ███╗ ██████╗ ███╗   ██╗████████╗ █████╗      ██╗ ██████╗   v2.6 · LOCAL  ║
║  ████╗ ████║██╔═══██╗████╗  ██║╚══██╔══╝██╔══██╗     ██║██╔═══██╗  EDIT-SUITE    ║
║  ██╔████╔██║██║   ██║██╔██╗ ██║   ██║   ███████║     ██║██║   ██║  #121118       ║
║  ██║╚██╔╝██║██║   ██║██║╚██╗██║   ██║   ██╔══██║██   ██║██║   ██║  #F4B33D       ║
║  ██║ ╚═╝ ██║╚██████╔╝██║ ╚████║   ██║   ██║  ██║╚█████╔╝╚██████╔╝                ║
╚══════════════════════════════════════════════════════════════════════════════════╝

  [SCRIPT] ──► [WHISPER] ──► [KOKORO-28] ──► [REMOTION] ──► [1.2s MP4]

  ├─ whisper-base     [██████████] 0.45× RT · word-level timestamps
  ├─ kokoro-v1.0      [████████░░] 28 Apache 2.0 voices · af_heart
  ├─ blender-5.2      [████████░░] port 8790 · WoL GPU stage worker
  └─ ffmpeg-concat    [██████████] bitstream copy · 3/4 scenes cached

  $ montajo render --delta scene-03 ⠋ READY  ▁▃▅▇▆▄▂▁  concat 1.14s · 0 untouched re-encodes █

Mission & Creator Proof

Built for High-Velocity Storytellers, Not Cloud Queues

Most AI video tools lock your footage behind slow cloud renders and per-minute transcription fees. Montajo puts a complete edit-suite on your own hardware—combining deterministic scene caching, word-timed motion graphics, and zero-lock-in exports.

Compute Efficiency 95% Saved

Only the edited scene slice re-renders; untouched clips concatenate via FFmpeg bitstream copy.

Local Transcription 0.45× RT

Faster-Whisper Base aligns word-level caption timestamps locally at zero recurring API cost.

Layout Reliability 170+

Stress-tested Remotion scene layouts across both 9:16 portrait and 16:9 landscape frames.

Subtitle Precision 144

Verified Unbounded & Figtree caption combinations with safe zones above UI overlays.

1. Edit in Seconds, Not Minutes

Fix a typo in scene 3 or swap a caption preset without waiting for a 60-second full timeline re-encode. Every scene hashes independently for instant 1–2s updates.

2. Word-Synced Motion & 3D

Charts, big stat counters, and flow diagrams trigger on the exact spoken word from Kokoro TTS, paired with headless Blender 5.2 3D stage renders.

3. Local-First & Self-Hostable

Run Montajo on your own workstation or Proxmox container with self-hosted Supabase GoTrue auth, plain MP4 + JSON outputs, and Apache 2.0 voices.

Dedicated Workflows

The Three Montajo Workspaces

Designed to match how modern video teams operate. Switch workspaces instantly with their own color themes, purpose-built tools, and independent project pipelines.

Studio Workspace

From Raw Topic to Polished Video in Minutes

Enter a topic or notes, choose portrait (9:16) or landscape (16:9), and let Claude structure an editorial scene plan. Montajo searches free stock footage (Pexels, Pixabay, Coverr) with 24h smart caching, narrates with Kokoro’s Apache 2.0 voice engine, generates Remotion charts and Blender 3D scenes, and compiles with word-highlighted captions.

✓
Free Stock Footage Auto-Selection: Pexels, Pixabay, and Coverr integrations with thumbnail swapper and creator attribution export.
✓
Kokoro 28 English Voices: 82M high-fidelity TTS model running locally at full precision (af_heart, am_michael, bf_emma).
✓
Automated 3-Day Topic Research: Claude scans live web sources every 3 days, verifies links, drops stale news (>14 days), and keeps an 8-idea pool.
✓
Remotion & Blender 3D Stage: Spring physics motion graphics landing on word cues, plus 3D titles and big numbers on a teal/magenta stage.

Try Studio Simulator python scripts/download_kokoro.py

studio_project_082/scenes.json

FRESH NEWS (2d old) The Verge · TechCrunch

AI Video Architecture: Why Scene Caching Beats Full Re-Renders

Angle: How modern short-form studios save 95% compute using incremental scene graphs.

3 verified sources 9:16 Portrait 4 Scenes (~16s)

1 Script Generated

2 Pexels Clips Picked

3 Kokoro Audio Synced

4 Remotion Cues Built

Campaigns Workspace

PPV Clipping Campaigns at Scale

Designed for creators running pay-per-view affiliate clipping campaigns. Drop a client brief (PDF, Word, or Google Doc link) and watch Claude parse the required rules, banned phrases, lengths, rate cards, and mandatory disclosure tags.

✓
Brief & Asset Auto-Sorter: Drop folders of raw videos, logos, watermarks, and music. Montajo automatically classifies roles.
✓
Local Whisper Transcription: High-precision word timestamps with zero API costs, running locally on your hardware.
✓
Claude 15–60s Moment Hunter: Scans transcript for the highest retention hooks that comply with campaign constraints.
✓
Auto 9:16 Reframe & Payout Log: Blurred-background fit or center-crop reframing, export with ready-to-paste tags, and clip approval log.

See Campaign Examples clip.cli video.mp4 --tags "#ad"

campaigns_brief_parser.py

CAMPAIGN: Epic Quest Launch · 9:16 Cutdowns

RATE: $1.50 per 1,000 verified views

MANDATORY TAGS: #EpicQuest #ad #gamingclip

DISCLOSURE: "Paid Partnership" tag enabled

CLIP_01.mp4 Score: 94/100 APPROVED ($450 est)

CLIP_02.mp4 Score: 88/100 POSTED (124K views)

Channel Workspace

YouTube Packaging & Cross-Platform Syndication

Take your long-form videos and turn them into an entire publishing bundle. Montajo writes high-CTR YouTube titles, chapters, SEO descriptions with stock credits, and tailors ready-to-post copy for TikTok, Instagram Reels, LinkedIn, and X.

✓
Full YouTube Packaging: Title variants, chapter timestamps, tags, and automatically formatted creator credits for stock footage.
✓
Per-Platform Tailored Copy: Formats hashtags, link disclosures, and character counts for TikTok, X, and Instagram.
✓
Thumbnail Concepts & Zip Exports: Automatically generates high-engagement thumbnail concepts and bundles clips in one zip.

See Channel Outputs Packaging: All Platforms

channel_packaging.json

YouTube TikTok Instagram Reels X / Twitter

TITLE A (HIGH CTR):

Why We Stopped Re-Encoding Videos: The 1-Second NLE

CHAPTERS:

00:00 The Re-Render Bottleneck · 02:15 Scene Graphs · 05:40 Benchmark

STOCK FOOTAGE CREDITS:

Visuals by Cottonbro Studio via Pexels (Free licence, included)