Auto News Video: Best AI Video Automation for Developers in 2026

What Is Auto News Video?

Auto News Video is one of the best AI Video Automation tools for Vietnamese tech content creators, indie hackers, and devrel teams. Built by hoquanghai, it turns a single article URL into a TikTok-ready 9:16 video in about 60 seconds of runtime once the pipeline is configured, with the README snapshot showing a 5-minute end-to-end workflow instead of a manual 3-hour edit.

The project is a code-first video factory, not a drag-and-drop editor. It generates script.json, validates it with Zod, renders with HyperFrames, mixes voice and SFX, and produces output/<slug>/video.mp4 plus sidecar assets for social posting and captions.

Quick Overview

Who Should Use Auto News Video?

• Vietnamese publishers and tech newsletters that need to turn written analysis into short-form video without hiring a full editor.
• Indie hackers shipping product updates on TikTok, Shorts, or Reels who want a repeatable render pipeline instead of ad hoc editing sessions.
• Devrel and marketing teams that need consistent branding, scene templates, and voice timing across a backlog of article-to-video conversions.
• Automation-heavy creators who are comfortable with the terminal, environment variables, and API keys for TTS and image generation.

Not ideal for:

• Teams that need frame-by-frame manual editing, keyframing, and timeline control in a traditional NLE like Premiere Pro.
• Creators who do not publish Vietnamese tech content and do not care about Vietnamese voice cloning or article ingestion.
• Users who want a hosted no-code service with support SLAs instead of a self-hosted Node.js pipeline.

Key Features of Auto News Video

• URL-to-video pipeline — Auto News Video accepts a URL, .txt, or .md input and turns it into a renderable script.json through a Claude Code slash command: /create-news-video <url>. That keeps the workflow close to the terminal and removes the need for manual scene planning.
• Twelve scene templates — The project ships with hook, comparison, stat-hero, feature-list, callout, outro, quote-card, icon-grid, timeline, big-text, chart-bars, and kinetic-quote templates. That template set covers news hooks, stats, comparisons, and punchy social-first layouts without rewriting the render engine.
• Multi-provider TTS — Auto News Video supports LucyLab for Vietnamese cloning and ElevenLabs for multilingual speech. The README also calls out idempotent synthesis, which means cached voice files are reused on re-renders instead of spending API quota again.
• Deterministic render stack — The core render path uses Puppeteer, GSAP, and FFmpeg through HyperFrames. That means the same input produces the same frames, which is the difference between a scriptable media pipeline and a fragile manual export session.
• Quality gates before render — lint, validate, and inspect run ahead of the final encode. The checks catch contrast failures, overflow, and off-canvas text before the MP4 is generated, which matters when the output is going straight to mobile feeds.
• Voice-text synchronization — Each scene can define voiceChunks, and beats fire when narration mentions the matching element. That reduces the common mismatch where a visual appears late because the editor guessed the timing by ear.
• Auto thumbnail generation — A Gemini 2.5 Flash Image cover is generated in 9:16 format and embedded into the MP4 without a separate re-encode pass. That saves time and keeps the asset pipeline simple when the video is pushed to Shorts or Reels.

Auto News Video vs Alternatives

Pick CapCut when you already have assets and want to trim, caption, and publish by hand. Auto News Video is better when the source is an article and you want the pipeline to do the repetitive work.

Pick HeyGen when your brief is an avatar or talking-head video. Pick manim-web when the video should look like a developer-authored explainer rather than a news reel. If your team wants a stronger LLM editing surface before rendering, Claude Code Canvas pairs well with Auto News Video for script shaping, while Moonshine Voice is a closer fit if you only need the narration layer.

How Auto News Video Works

Auto News Video splits the job into two halves: AI-generated content and deterministic rendering. Claude Code reads the source article, extracts a structure, and emits script.json; then Node.js, TypeScript, and HyperFrames turn that schema into frames, audio, and final video output. That design keeps the creative step flexible while keeping the render step reproducible.

The data model matters here. The repository uses Zod discriminated unions for the 12 template variants, which makes invalid scene payloads fail before render time. The pipeline also carries per-scene voiceChunks, SFX rules, and quality checks, so a scene is not just text but a typed bundle of narration, motion, and layout constraints.

# getting started example
claude
/create-news-video https://vnexpress.net/some-article
# or, without Claude Code
npm run pipeline -- output/my-video/script.json

The first command asks Claude Code to fetch the article, draft the script, and hand the structure to the renderer. The second command skips the assistant layer and runs the pipeline directly from a hand-authored JSON file, which is useful when you want deterministic inputs or are testing a specific scene template.

Pros and Cons of Auto News Video

Pros:

• Fast article-to-video conversion — The README claims a 5-minute workflow instead of roughly 3 hours of manual scripting, voice work, and editing.
• Type-safe render pipeline — Zod validation and TypeScript ESM reduce schema drift and catch bad scene data before FFmpeg starts encoding.
• Built-in Vietnamese voice workflow — LucyLab support gives the project a real advantage for Vietnamese narration, especially if you need cloning.
• Template variety — Twelve templates cover most short-form news and product-update formats without custom animation work.
• Idempotent audio generation — Reused voice files cut API spend when you regenerate the same script with minor layout changes.
• Production-style checks — Contrast validation and overflow inspection are the kind of preflight steps that keep mobile videos readable.

Cons:

• API dependencies are mandatory — You still need TTS and image-generation keys, so the pipeline is not fully offline.
• Narrow content fit — It is optimized for Vietnamese tech articles, not arbitrary long-form essays, interviews, or live-action storytelling.
• Learning curve for non-engineers — The setup assumes comfort with npm, .env files, and a CLI-first workflow.
• Render stack complexity — HyperFrames, Puppeteer, GSAP, FFmpeg, and external APIs are a lot to debug when something breaks.
• Not a full editor — Auto News Video does not replace Premiere Pro or a timeline editor when you need manual motion design or frame-accurate post production.

Getting Started with Auto News Video

Clone the repo, install dependencies, and set your API keys before running the first render. The quickest path is the Claude Code route because it hides the script authoring step behind a slash command.

git clone https://github.com/hoquanghai/Auto-Create-Video.git
cd Auto-Create-Video
npm install
cp .env.example .env.local
# edit .env.local and set TTS_PROVIDER plus your LucyLab or ElevenLabs key

# recommended path
npm install -g @anthropic-ai/claude-code
claude
/create-news-video https://vnexpress.net/some-article

After that, the pipeline writes its artifacts into output/<slug>/, including video.mp4, voice.mp3, script.txt, and the social caption file. If you skip Claude Code, you can hand-write script.json from src/render/script-schema.ts and run npm run pipeline -- output/my-video/script.json directly.

Verdict

Auto News Video is the strongest option for turning Vietnamese tech articles into short-form video when you want a code-driven pipeline instead of manual editing. Its best trait is the deterministic Node/TypeScript render stack, and its main caveat is the dependency on external AI APIs. If your workflow starts from article URLs, this is the right choice.