# [1.14.0](https://github.com/stirling-image/stirling-image/compare/v1.13.0...v1.14.0) (2026-04-10) ### Bug Fixes * add FILES_STORAGE_PATH to Dockerfile ENV to prevent data loss ([b575243](https://github.com/stirling-image/stirling-image/commit/b575243e9a0cb2ac4567d785a74e57dea912e9e2)) * add shutdown timeout and improve health endpoint ([986ad37](https://github.com/stirling-image/stirling-image/commit/986ad37bb5e6644dd93521018a1bcd2d6243f502)) * address code review findings before merge ([caf65bc](https://github.com/stirling-image/stirling-image/commit/caf65bc4697ddfea27674b56fd2e8de847edc734)) * correct PaddleOCR language codes for model download and OCR ([e1ee571](https://github.com/stirling-image/stirling-image/commit/e1ee57103c201f104b737c2719d0014d4912d4b2)) * force CPU mode in download_models.py for build-time compatibility ([b4b59a7](https://github.com/stirling-image/stirling-image/commit/b4b59a7500375bc2e65cb0050e5308b07610d2cf)) * handle paddlepaddle-gpu CUDA import at build time gracefully ([0083a74](https://github.com/stirling-image/stirling-image/commit/0083a741a9a3924ccc37c4e622c2980a8a41f1a3)) * install cuda-compat stubs for build-time PaddlePaddle import ([d31d665](https://github.com/stirling-image/stirling-image/commit/d31d66556ef0aa7c20a73a94f5d95b54b99fbb8a)) * load RealESRGAN pretrained weights for actual AI upscaling ([fa9569c](https://github.com/stirling-image/stirling-image/commit/fa9569c920d6bdac094dcd661fc5cf3b4be3f17a)) * revert to npx tsx in CMD for pnpm compatibility ([e55253d](https://github.com/stirling-image/stirling-image/commit/e55253dee03f9e93187d50e0ffb66b423a1230c0)) * simplify smoke test to CPU-only imports for build-time compat ([3481663](https://github.com/stirling-image/stirling-image/commit/34816639609cbccab91a5cad5760fea769c4b565)) * skip RealESRGAN import check on arm64 in smoke test ([1e2ef52](https://github.com/stirling-image/stirling-image/commit/1e2ef5284686253ae33e05411a515369b7f41a3e)) * split paddlepaddle-gpu and paddleocr installs, use --extra-index-url ([74183e8](https://github.com/stirling-image/stirling-image/commit/74183e8dc1e3d8f90cff4484f5526c3c3dfba9a6)) * suppress ML library stdout noise in ocr.py and upscale.py ([c0b419d](https://github.com/stirling-image/stirling-image/commit/c0b419de21acb92b243ef78191dff6a2153c5961)) * use PaddlePaddle GPU package index for CUDA wheels ([dd9528f](https://github.com/stirling-image/stirling-image/commit/dd9528f53c9bcf148842ab94f14a4ebd5ae223ec)) * use platform-specific mediapipe version for arm64 compatibility ([7face19](https://github.com/stirling-image/stirling-image/commit/7face19238f54cfea02d453b697e34b50ccd9d21)) ### Features * expand model pre-download with verification and smoke test ([a9e3b96](https://github.com/stirling-image/stirling-image/commit/a9e3b9688776a51efd2f42cae810de425cf1d9cf)) * simplify CI to single unified Docker build ([b385a2e](https://github.com/stirling-image/stirling-image/commit/b385a2eabb255f7a8fe2d3720a358ff5bf254310)) * simplify compose to single file, add log rotation ([84f7057](https://github.com/stirling-image/stirling-image/commit/84f7057a49850b7a47440099ec2caad8d3b87efe)) * unified Docker image with GPU auto-detection ([6c3eb3b](https://github.com/stirling-image/stirling-image/commit/6c3eb3b876cee0301d5fb3ed8324a3c8e92b1307))
We're renaming! The Stirling PDF team asked us to pick a new name, and we want you to choose it.
Vote on the new name →
Voting closes April 10, 2026
Stirling Image
Stirling-PDF but for images. 30+ tools and local AI in a single Docker container.
Key Features
- 30+ image tools - Resize, crop, compress, convert, watermark, color adjust, and more
- Local AI - Remove backgrounds, upscale images, erase objects, blur faces, extract text (OCR). All running on your hardware with pre-downloaded models, no internet required
- Pipelines - Chain tools into reusable workflows. Batch process up to 200 images at once
- REST API - Every tool available via API. Interactive docs included at
/api/docs - Single container - One
docker run, no Redis, no Postgres, no external services - Multi-arch - Runs on AMD64 and ARM64 (Intel, Apple Silicon, Raspberry Pi)
- Your data stays yours - No telemetry, no tracking, no external calls. Images never leave your machine
Quick Start
docker run -d -p 1349:1349 -v stirling-data:/data stirlingimage/stirling-image:latest
Open http://localhost:1349 in your browser.
Looking for the Lite or CUDA image? Click here.
Lite (~1.5 GB) - all image tools, no AI:
docker run -d -p 1349:1349 -v stirling-data:/data stirlingimage/stirling-image:lite
CUDA (~14 GB) - GPU-accelerated background removal, upscaling, and OCR:
docker run -d -p 1349:1349 --gpus all -v stirling-data:/data stirlingimage/stirling-image:cuda
CUDA needs an NVIDIA GPU and Container Toolkit. Falls back to CPU if no GPU is found. See Docker Tags for benchmarks and Docker Compose examples.
Default credentials:
| Field | Value |
|---|---|
| Username | admin |
| Password | admin |
You will be asked to change your password on first login. This is enforced for all new accounts and cannot be skipped in production.
For Docker Compose, persistent storage, and other setup options, see the Getting Started Guide. For details on all image variants (full, lite, cuda), see Docker Tags.
Documentation
Contributing
Contributions welcome. See CONTRIBUTING.md for guidelines, the Developer Guide for setup, and the Translation Guide for adding languages.
Support
Bug reports and feature requests: GitHub Issues
License
This project is dual-licensed under the AGPLv3 and a commercial license.
- AGPLv3 (free): You may use, modify, and distribute this software under the AGPLv3. If you run a modified version as a network service, you must make your source code available under the AGPLv3. This applies to personal use, open-source projects, and any use that complies with AGPLv3 terms.
- Commercial license (paid): If you want to use Stirling Image in proprietary software or SaaS without the AGPLv3 source-disclosure requirement, a commercial license is available. Contact me for pricing and terms.

