Generate seamlessly looping animated movie posters from still images using AI image-to-video models via PiAPI.
Wonder Woman • Back to the Future • Pacific Rim • Avatar: Fire and Ash • Stargate • Jurassic Park
All generated with Kling AI via PostAnimate
Point it at a movie-poster image and a short motion prompt, and PostAnimate prepares the image, runs it through one of 11 AI image-to-video models via PiAPI's unified API, and post-processes the result into a seamlessly looping video clip. Use it as a CLI or as a Python library.
- 11 AI models — Kling, Luma, Hunyuan, Hailuo, Wan, Wan 2.6, Veo 3, Veo 3.1, SkyReels, FramePack, Seedance
- Seamless looping — crossfade blending or boomerang playback for a smooth loop point
- Poster-native aspect ratio — preserves the exact 27:40 movie poster ratio (1296 × 1920)
- Automatic image prep — resizes, pads, or crops any image to the correct dimensions
- Unified API — every supported model accessed through one PiAPI account and one API key
- Library + CLI — use as a Python package or from the command line
git clone <this-repo>
cd animated_poster_generation
pip install -r requirements.txt
cp .env.example .env # add your PiAPI key
python postanimate_cli.py animate input/poster.jpg \
-p "subtle floating particles, gentle light flickers"Output lands in output/:
<name>_prepared.jpg— resized poster sent to the API<name>_raw.mp4— the model's raw output<name>_loop.mp4— the final seamlessly looping clip
Python 3.10+ and FFmpeg on PATH are required. Image uploads need the PiAPI Creator plan or higher.
| Doc | What's in it |
|---|---|
| Setup | Install dependencies, FFmpeg, API key, environment variables. |
| Models | All 11 models — comparison table, deep-dive on each, "which model should I use?". |
| CLI reference | Every subcommand (animate, models, loop, prep, fetch, extend) with every flag, plus a big examples block. |
| Library usage | Using postanimate as a Python package — full pipeline, individual components, batch. |
| Prompt guide | How to write motion prompts, a copy-paste LLM meta-prompt, and a ~35-prompt ready-made library organized by element. |
| Internals | Pipeline walkthrough, loop modes, project structure, adding new models. |
| Troubleshooting | Common errors and their fixes. |
MIT