Caption-Creator Lands: Local Image Captioning Skips the Cloud
Caption-Creator is a fast, portable GUI tool that runs entirely on your local machine to generate high‑quality image captions and tags. It helps users build custom datasets for AI model training or reverse‑engineer prompts from existing images. The application works without any installation, keeping all processing private and uncensored.
Merserk developed this open‑source project to simplify dataset creation, particularly for LoRA training and advanced image prompting. He focused on making the tool fully self‑contained, eliminating complex setup or Python environment requirements. Because everything runs locally, there are no content filters and user data never touches cloud servers.
Dual generation modes and local control
Key Features
- Switch between detailed captions and comma tags.
- Process whole folders with a gallery‑style view.
- Paste images directly from the clipboard.
- Pick models optimized for 6 GB to 10+ GB VRAM.
- Connect to LM Studio for any compatible model.
- Keep model in VRAM to speed up later runs.
The tool suits users with consumer GPUs who need to caption large image collections locally. Small agencies can produce training data without relying on cloud services or dealing with content filters. Privacy‑conscious professionals will appreciate that no data ever leaves their machine.
Technical considerations and community notes
Merserk cautions that enabling low‑VRAM mode may fit more model layers into GPU memory but can significantly slow generation speed. The application is distributed as a single portable folder with embedded Python, so no external dependencies are needed. While no detailed roadmap is provided, the tool already supports interactive model management and future updates could expand on this.
"The application runs entirely on your local machine, ensuring privacy and uncensored output." — Source: GitHub
Project Links
