PTENES
AI video production · local

From content to video ready

Create a script, voice, images, animations, captions, and an MP4 in a visual workflow that lets you review each step.

Start the guideView on GitHub
Visual production line connecting the scenes of a video
What it is

A local, verifiable studio

VideoSub turns a short topic or complete content into a vertical, square, or horizontal video, keeping you in control of the result.

🧭 Visible steps

Script, voice, image, motion, captions, and render are shown separately for review.

🔌 Real providers

Codex OAuth creates the script, ElevenLabs produces the voice, and Agnes generates images and videos.

🏠 Local execution

The interface and API run on your machine, with SQLite for projects and FFmpeg for final assembly.

How it works

A seven-step production line

Each output is preserved in the project and can be reviewed before the next operation.

Content→Script→Voice→Images→Animation→Captions→MP4
Prerequisites

What you need to run it

The application uses local tools and credentials configured in an untracked file.

Node.js and pnpm

Node.js 22 or later and pnpm 10.

node --version
pnpm --version

FFmpeg

FFmpeg and FFprobe available on the PATH, including the subtitles filter.

ffmpeg -version
ffprobe -version

Providers

Codex authenticated, with Agnes and ElevenLabs keys in the local environment.

codex login status
# configure apps/server/.env
User guide · step by step

Run it and produce the first video

The commands below are the same ones used in the local version of the project.

1

Install the dependencies

At the repository root, install the workspace.

pnpm install --frozen-lockfile
2

Create the local configuration

Copy the template and fill in the credentials without versioning the file.

cp .env.example apps/server/.env
nano apps/server/.env
3

Build and start

The script creates the build when needed and serves the interface on port 3333.

pnpm build
./start.sh
# open http://localhost:3333
4

Create a project

Enter a topic or content, choose a format and style, and generate the script.

Tema: explique o que é AGI
Formato: 9:16
Estilo: editorial cinematográfico
5

Validate each stage

Generate voice, images, and animations; review the scenes, then create captions and render.

# interface workflow
Roteiro → Voz → Imagens → Animar → Legendas → Renderizar
6

Update or stop

Use the operational scripts included in the repository.

./atualiza.sh  # validate, build, and restart
./stop.sh      # stop the server
Real example

Explainer video about AGI

Frames from the end-to-end validated result with ElevenLabs voice, Agnes images and motion, and local editing.

Vertical scene with an artificial brain
Cinematic opening generated to explain the concept of AGI.
Vertical scene about the impacts of artificial intelligence
Final scene with compact captions and a blue-and-amber composition.
Roadmap

From MVP to complete product

The current setup prioritizes real production and local validation; the next phases expand operations and polish.

CURRENT
Real local pipelineScript, voice, images, video, captions, and final render with provider diagnostics.
NEXT
Creative controlsMore styles, voices, music mixing, caption templates, and per-scene adjustments.
FUTURE
Multiuser operationAuthentication, job queue, external storage, publishing, and cost tracking.