lcy362/agnes-video-generator

Agnes Video Generator – Free, Open‑Source AI Video Creator

What it is – A Python‑based application that lets anyone turn text (or images) into multi‑scene videos with AI‑generated narration, subtitles, and optional digital‑anchor avatars. It runs the heavy AI compute in the cloud via the free Agnes AI API, so you only need a normal laptop.

Key capabilities

• Text‑to‑video & manuscript mode – Paste a story or article; the tool splits it into segments, generates a short clip for each, then stitches them together with TTS narration and word‑level SRT subtitles.
• Image‑to‑video & keyframe animation – Provide a sequence of images or keyframes; the system animates them into a coherent clip.
• Digital anchor – Insert a synthetic presenter that “talks” the narration.
• Auto subtitles – Generated subtitles are exported as SRT files.
• No GPU required locally – All model inference happens on Agnes AI’s cloud service; the client only talks to a REST API.
• Unlimited usage – Free tier has no hard usage caps (only a 16‑requests‑per‑minute rate limit per API key). Multiple keys can be chained for higher throughput.

How to get started

1. Obtain an API key from the free Agnes AI platform.
2. Clone the repo and copy the example environment file:

   git clone https://github.com/lcy362/agnes-video-generator.git
   cd agnes-video-generator
   cp .env.example .env   # edit AGNES_API_KEY (and optional extra keys)
   

3. Run the server in one of the supported ways:
   • Manual – ./start.sh (Python 3.10+ required).
   • Docker – docker pull lcy362/free-short-video then docker run ….
   • npm – npx free-short-video (wraps the Docker image).
   • AI‑agent assisted – provide the AGENTS.md file to a coding assistant; the assistant can install, configure, and launch the service automatically.
4. Open the web UI (default http://localhost:8000) or call the REST API directly (see docs/public/api.md).

Typical workflow

• Choose a mode (Creative, Manuscript, Image‑to‑Video, etc.).
• Enter your prompt or upload images.
• The backend calls the Agnes AI video model, then runs TTS and subtitle generation.
• When finished you receive a downloadable MP4 and an SRT file.

Project health & community

• MIT‑licensed, fully open source.
• CI badge shows SonarCloud quality gate passed and test coverage.
• Docker Hub image (lcy362/free-short-video) is publicly pulled.
• The author provides a companion skill‑set (agnes-video-generator-skills) and documentation for AI‑agent operation, encouraging automated deployment and debugging.

Why it matters Commercial AI video platforms (Runway, Pika, etc.) charge monthly fees or per‑second costs and often require a paid subscription. Agnes Video Generator offers the same multi‑scene, narrated video pipeline for free, with no watermark and no usage limits, making AI video creation accessible to hobbyists, educators, and small creators.

Links

• Live demo / web UI – https://video.lichuanyang.top (Simple Video mode)
• Documentation – docs/public/ (features, getting‑started, API reference, agent operations)
• Related tools – FreeShortVideoStudio (browser‑only UI), Flint (skill manager for AI agents)

---

All information above is taken directly from the repository’s README; no additional features are inferred.