Pixelle-Video ATH-MaaS/Pixelle-Video

Topic-to-short-video engine that writes the script, generates AI images or clips through ComfyUI or APIs, adds voice and music.

Creator kit + Compare GitHub ↗ Install guide (official) ↗

Pixelle-Video at a glance

Checked 16 Oct 2026

Pixelle-Video (Apache-2.0) turns a topic into a narrated short video: LLM script, an AI image or clip per sentence via ComfyUI workflows (local or RunningHub cloud) or direct image/video APIs, Edge-TTS or Index-TTS voice, music and templates. Extra pipelines cover a digital-human avatar, image-to-video and motion transfer. The README states a zero-cost setup with Ollama plus local ComfyUI, which needs your own GPU; otherwise it uses paid APIs. Windows all-in-one package; uv install on macOS/Linux. Hardware needs are not documented.

Cite this pageRepoQuarry. “Pixelle-Video.” Checked 16 Oct 2026. https://repoquarry.dev/projects/pixelle-video

How much of InVideo AI does it replace?

4 of 5 jobs. Generates the visuals with AI per sentence instead of stock footage; free only with local Ollama and ComfyUI on your own GPU.

Script or prompt to video Partial replacement

Faceless short videos from a topic Partial replacement

AI voiceover Partial replacement

Automatic subtitles Not covered

AI-generated images and video clips Partial replacement

For automatic subtitles, see other InVideo AI alternatives. You may also need to build an interface, an API server and hosting yourself.

How much of HeyGen does it replace?

1 of 3 jobs. Has a Digital Human pipeline (added January 2026); not tested against HeyGen avatars.

AI avatar (talking presenter) videos Partial replacement

Video translation and dubbing Not covered

Script or prompt to video Not covered

For video translation and dubbing, see other HeyGen alternatives. You may also need to build an interface, an API server and hosting yourself.

Facts, with sources

Every answer links to where we found it. We read the documentation; we did not install it.

Report a correction

QuestionAnswerSourceChecked

Runs on my machine? Yes, local web UI on localhost:8501 Browser will automatically open http://localhost:8501 github.com · README_EN.md#-quick-start ↗ 1d ago

Code licence Apache-2.0 This project is released under the Apache License 2.0. github.com · LICENSE ↗ 1d ago

Model weights licence No single bundled model licence: users select Ollama, ComfyUI/RunningHub workflows or direct model APIs. Each selected weight or provider keeps its own licence and service terms; Apache-2.0 covers the application code Each step supports flexible customization, allowing you to choose different AI models github.com · README_EN.md ↗ today

Commercial use Apache-2.0 code; check the licence of each image/video model or API you use github.com · LICENSE ↗ 1d ago

Paid services needed? Not required with Ollama + local ComfyUI (needs a capable GPU); otherwise paid LLM, RunningHub or image/video APIs Completely Free Solution: LLM using Ollama (local) + ComfyUI local deployment = 0 cost github.com · README_EN.md ↗ 1d ago

Hardware Not documented; the README recommends the free local route only if you have a local GPU If you have a local GPU, recommend completely free solution, otherwise recommend using Qwen (cost-effective) github.com · README_EN.md ↗ 1d ago

Operating systems Windows (all-in-one package), macOS and Linux (from source) Install from Source (For macOS / Linux Users or Users Who Need Customization) github.com · README_EN.md#-quick-start ↗ 1d ago

Setup effort Easy on Windows (all-in-one package); uv + ffmpeg on macOS/Linux; local ComfyUI is extra work No need to install Python, uv, or ffmpeg - ready to use out of the box! github.com · README_EN.md#-quick-start ↗ 1d ago

What it costs to run

Software$0

Required paid servicesnone

Optional paid servicesOpenAI, Qwen or other LLM APIs, RunningHub cloud ComfyUI, DashScope, Seedance, Kling and other media APIs

Your hardwareNot documented; the README recommends the free local route only if you have a local GPU

InVideo AI: Paid plans per seat: Starter $10/month, Plus $30/month, Max $75/month billed annually (shown as 50% off); monthly billing costs more (source, checked 15 Oct 2026). Running it yourself costs your hardware, electricity and setup time.

Project activity

29kstars

4.2kforks

18contributors

Last code change4mo ago

Latest releasev0.1.15 · 8mo ago

ArchivedNo

From GitHub, today. Stars show popularity, not quality.

Using an AI agent?

The same record, as JSON or through MCP.

get_project("pixelle-video")
GET /api/v1/projects/pixelle-video.json

API and MCP docs →