Turn Any Topic Into a Studio-Quality Video
DocuForge AI scrapes live data from 10+ sources, synthesizes research with Claude, writes scripts with Llama 3.1, generates voice with ElevenLabs, creates visuals with Flux, and renders cinematic MP4s — zero editing required.
- Viral 9:16 shorts for TikTok, Reels & YouTube Shorts
- 21:9 cinematic documentaries in Vox & BBC style
- Burned-in subtitles with word-level timestamps
- Admin dashboard to control all API keys & scrapers
No credit card required · Cancel anytime
● Generating Video
Viral Short · 60sTopic
“Why Japan's population is collapsing — and what it means for the world”
Research
Tavily · Jina · Serper · Exa · Wikipedia — 8/10 sources
Claude Brief
14 key facts · 3 hook angles · timeline built
Script
Llama 3.1 — 12 scenes · 58 sec total
Assets + Render
Flux images · ElevenLabs · Remotion 9:16
10+
Live Data Sources
Tavily, Jina, Serper, Exa & more
50K
Chars Researched
Per video generation job
2
Output Formats
9:16 viral + 21:9 cinematic
~45m
Avg. Production
Topic to finished MP4
Powered by industry-leading AI & data infrastructure
How It Works
From Topic to Published Video
Four automated phases handle everything — you just enter a topic and download the MP4. No editing software, no film crew, no stock footage subscriptions.
30 seconds
Describe Your Topic
Type any subject into the Vidrush-style create studio. Choose Viral Short (9:16), Documentary (21:9), or Listicle format. Set duration from 15 seconds to 40 minutes.
- Natural language topic input
- Format & duration selector
- Optional custom script paste
- Voice & style preferences
5–15 minutes
AI Researches & Scripts
10+ scrapers fetch live data in parallel. Claude AI synthesizes a research brief. Llama 3.1 or Claude writes a scene-by-scene script with narration, image prompts, and map coordinates.
- Tavily, Jina, Serper, Exa, NewsAPI scraping
- Claude research brief synthesis
- Chronological scene JSON output
- Hook-driven viral script structure
10–25 minutes
Assets Are Generated
ElevenLabs narrates each scene with word-level timestamps. Flux generates photorealistic vertical images. Wikimedia, Pexels, and Pixabay provide real-world B-roll and archival media.
- ElevenLabs multilingual voice
- Flux 1-dev image generation
- Stock & archival media fetch
- Wan2.1 image-to-video animation
5–15 minutes
Video Is Rendered
Remotion assembles scene clips with Motion.dev transitions. Mapbox renders animated map sequences. FFmpeg burns subtitles, applies audio ducking, and exports final MP4.
- Remotion + Motion.dev assembly
- Center-bottom subtitle burn-in
- Audio ducking & transition SFX
- 9:16 or 21:9 MP4 download
Technology
6-Phase Production Pipeline
Every video passes through research, intelligence, voice, visuals, character rendering, and final assembly — fully automated, production-grade output.
01 — Research
Multi-Source Live Scraping
10 scrapers run in parallel, collecting up to 50,000 characters of live research. Admin dashboard lets you enable/disable each source and configure API keys independently.
02 — Intelligence
Claude + Llama Script Engine
Claude AI synthesizes all scraped data into a research brief with key facts, hook angles, and timeline. Llama or Claude then writes scene-by-scene scripts with image prompts, narration, and map coordinates.
03 — Voice
ElevenLabs Narration
Professional AI voice synthesis with character-level word timestamps for precise subtitle burn-in. Supports 29+ languages. Audio ducking drops background music 85% during narration.
04 — Visuals
Flux + Stock Media
Abstract scenes get photorealistic Flux images (1080×1920 vertical). Real topics pull archival photos from Wikimedia and Internet Archive. B-roll from Pexels/Pixabay. Static images animated via Wan2.1.
05 — Characters
DeepVideo Character Engine
Historical character scenes use LivePortrait for lip-sync, then DeepVideo-V1 for micro-expressions, temporal consistency, and anti-flicker neural rendering at 1080p fidelity.
06 — Assembly
Remotion + FFmpeg Render
Remotion orchestrates scene clips with spring transitions. Animated geopolitical maps from coordinates. FFmpeg burns center-bottom subtitles, inserts whoosh/boom SFX, and exports 9:16 or 21:9 MP4.
Platform Features
Everything in One Production Pipeline
Nine integrated capabilities that replace an entire video production team — researchers, scriptwriters, voice actors, graphic designers, and editors.
10+ Live Data Scrapers
Parallel web research from Tavily, Jina AI, Serper, Firecrawl, Exa, Brave Search, NewsAPI, Google CSE, Wikipedia, and Internet Archive.
- Up to 50,000 chars per job
- Admin toggle per source
- Automatic failover
Dual AI Script Engine
Claude Sonnet synthesizes research briefs. Llama 3.1 70B (English) and Qwen 2.5 72B (Hindi/Urdu) write scene-by-scene viral scripts.
- Research brief with key facts
- Hook angle identification
- Chronological scene JSON
ElevenLabs Pro Narration
Multilingual v2 voice model with character-level word timestamps. Enables precise subtitle burn-in and audio ducking during narration.
- 29+ languages
- Word-level timestamps
- Custom voice ID support
Flux AI Visual Generation
Photorealistic 1080×1920 vertical images for abstract scenes. Wan2.1 animates static images into 4-second cinematic motion clips.
- NVIDIA NIM Flux 1-dev
- Vertical-optimized prompts
- Ken Burns + motion
Character Cinematics
LivePortrait delivers precise lip-sync for historical figures. DeepVideo-V1 adds micro-expressions, temporal consistency, and anti-flicker rendering.
- LivePortrait lip-sync
- DeepVideo-V1 enhancement
- 1080p neural rendering
Remotion Video Assembly
Remotion.dev orchestrates scene clips with Motion.dev spring transitions. Animated Mapbox map sequences from script coordinates.
- Procedural transitions
- Geopolitical map overlays
- Multi-layer timeline
Auto Subtitle Burn-In
FFmpeg processes ElevenLabs word timestamps into center-bottom aligned SRT subtitles, burned directly into the final MP4 export.
- Word-level sync
- Center-bottom alignment
- Karaoke-style highlight
Animated Map Sequences
Mapbox and OpenStreetMap integration renders dynamic zoom animations for geopolitical and historical documentary sequences.
- Lat/lng from script JSON
- Animated zoom transitions
- Location label overlays
Admin API Control
Supabase-powered settings table stores all API keys. Admin dashboard with scraper status grid, enable toggles, and 60-second cache refresh.
- Dynamic key management
- Per-scraper toggles
- No redeploy needed
Output Specs
Two Formats, Production-Grade Quality
Every export is a finished, publish-ready MP4 — no post-production editing required. Subtitles, transitions, voice, and visuals are baked in.
Viral Short
9:16 · 1080×1920 · 15 – 90 sec
Optimized for
- Hook-driven script structure
- Fast-paced scene cuts (3–5 sec)
- Center-bottom karaoke subtitles
- Flux vertical image generation
Cinematic Documentary
21:9 · 2560×1080 · 5 – 40 min
Optimized for
- Vox / BBC documentary pacing
- Animated Mapbox map sequences
- DeepVideo character cinematics
- Archival media from Wikimedia
Codec
H.264 MP4
Avg. render
~45 minutes
Languages
EN · HI · UR · Roman
Max file size
Up to 2 GB
Use Cases
Built for Every Type of Creator
Whether you run a YouTube channel, manage brand social accounts, or teach online — DocuForge adapts to your workflow with two distinct output formats.
YouTube Creators
Produce 2–3 documentary videos per week without a film crew. Vox-style explainers, history channels, and true-crime narratives with real archival footage and professional narration.
21:9 cinematic · 5–40 min
Social Media Marketers
Generate scroll-stopping 9:16 viral shorts from any trending topic. Claude researches live news, writes hook-driven scripts, and renders TikTok/Reels-ready MP4s with burned-in subtitles.
9:16 vertical · 15–90 sec
Educators & Course Creators
Turn complex subjects into engaging visual lessons. Multi-source Wikipedia, Serper, and Exa research ensures factual accuracy. Supports English, Hindi, Urdu via Qwen 2.5.
Multi-language · Subtitled
News & Media Outlets
Rapidly produce breaking news explainers with live data from NewsAPI, Tavily, and Jina. Get a fully narrated, visual video within minutes of a story breaking.
Live research · 10+ APIs
Agencies & Brands
White-label video production for clients at scale. Admin dashboard controls all API keys. Batch-generate branded content with consistent ElevenLabs voice and Flux visual style.
Admin API control · Pro plan
Indie Filmmakers
Prototype documentary concepts before investing in production. DeepVideo-V1 character cinematics, LivePortrait lip-sync, and Wan2.1 animation for historical figure scenes.
Character AI · Archival media
Comparison
Why DocuForge vs Alternatives
See how our end-to-end pipeline compares to manual editing and generic AI video tools.
| Feature | DocuForge | Manual | Other AI |
|---|---|---|---|
| Live web research (10+ sources) | |||
| Claude AI research synthesis | |||
| Automated scriptwriting | |||
| ElevenLabs voice + timestamps | |||
| Flux AI image generation | |||
| Character lip-sync (DeepVideo) | |||
| Burned-in subtitles | |||
| 9:16 + 21:9 output | |||
| Time to produce (avg) | ~45 min | 3–5 days | 2–4 hrs |
| Cost per video (Pro) | $4.83 | $500+ | $5–15 |
Testimonials
Trusted by Creators Worldwide
Join thousands of YouTubers, marketers, and educators who automate their video production with DocuForge AI.
“I went from spending 3 days per documentary to under an hour. The multi-source research alone saves me 6+ hours of manual fact-checking. My retention rate jumped 40% after switching to DocuForge scripts.”
Marcus Chen
History YouTuber · 280K subs
“We produce 15 viral shorts per week for our agency clients. The Claude research brief is incredibly detailed — clients think we have a full research team. ROI on the $29 Pro plan is insane.”
Sarah Okonkwo
Founder, ViralFrame Agency
“As an educator, factual accuracy matters. DocuForge pulling from Wikipedia, Serper, and NewsAPI simultaneously gives me confidence in every video I publish to my 50K student audience.”
Dr. Raj Patel
Online Course Creator
Pricing
Simple, Transparent Pricing
5 credits = 1 video. Free plan includes 1 video (5 credits). Pro gives 6 videos (30 credits) for $29/mo.
Starter
Try DocuForge with one complete video — no card required.
- 1 video render (5 credits)
- 5 credits deducted per video
- Viral Short 9:16 format
- English scripting (Llama 3.1)
- 5 scrapers (Tavily, Jina, Wikipedia + more)
- ElevenLabs standard voice
- 720p output with subtitles
- Community support
- —Documentary 21:9 format
- —Claude research brief
- —DeepVideo characters
- —Priority queue
Pro
Publish weekly — 6 full video renders every month.
- 6 video renders per month (30 credits)
- 5 credits per video render
- All formats: Viral Short + Documentary + Listicle
- All 10+ scrapers with parallel research
- Claude AI research brief synthesis
- Llama 3.1 + Qwen 2.5 (Hindi, Urdu, Roman)
- DeepVideo-V1 character cinematics
- 21:9 cinematic 1080p + 9:16 vertical
- Flux AI images + Wan2.1 animation
- ElevenLabs premium multilingual voices
- Priority rendering queue
- Commercial usage rights
- Admin API key dashboard
Each render costs 5 credits. Failed jobs restore credits automatically.7-day money-back guarantee on Pro.
FAQ
Common Questions
Everything you need to know about credits, formats, research sources, and commercial usage.
Get Started Today
Your Next Viral Video Is One Topic Away
Join creators who publish daily content without a production team. Enter a topic, wait ~45 minutes, download your MP4.
- 1 free video (5 credits) — no card required
- 5 credits per complete render
- Full pipeline: research → script → render
- Download MP4 with burned-in subtitles
- Cancel Pro subscription anytime