- Add replace-color tool with pixel-level color replacement and tolerance
- Register all 19 new Phase 3 tools in routes/tools/index.ts
- Map all new tool IDs to settings components in tool-page.tsx
- Add PDF, ZIP, ICO, JSON content types to download route
- Install qrcode, jsqr, potrace, pdfkit dependencies
Add 6 tools for format conversion and optimization extras:
- svg-to-raster: SVG to PNG/JPG/WebP at custom resolution
- vectorize: raster to SVG via potrace (B&W and color modes)
- gif-tools: animated GIF resize, frame extraction, optimization
- bulk-rename: pattern-based file renaming with ZIP output
- favicon: generate all favicon/app icon sizes with manifest.json
- image-to-pdf: combine images into PDF using pdfkit
Add 3 layout and composition tools with API routes and frontend settings:
- collage: multi-image grid layout with configurable gap and background
- split: image grid splitting with ZIP output (reuses archiver pattern)
- border: borders, rounded corners via SVG mask, padding, and shadows
Add 4 watermark/overlay tools with API routes and frontend settings:
- watermark-text: SVG text overlay with tiling, position, opacity, rotation
- watermark-image: logo/image watermark with position, opacity, scale
- text-overlay: styled text on images with shadow and background box
- compose: layer images with position, opacity, and blend modes
Backend: POST /api/v1/tools/:toolId/batch accepts multiple files +
settings, processes via p-queue with CONCURRENT_JOBS concurrency limit,
streams ZIP response using archiver. Tool registry in tool-factory
enables batch to reuse any registered tool's process function. SSE
endpoint at GET /api/v1/jobs/:jobId/progress provides real-time updates.
Handles partial failures gracefully, preserves filenames, deduplicates
collisions. Frontend: use-batch-processor hook handles upload, SSE
progress tracking, and automatic ZIP download.
Draggable split-view slider with clip-path reveals original vs processed
image. Shows file size badges with savings percentage. Supports pointer
events for both mouse and touch. Integrated into tool-page to replace
dropzone after processing completes. Updated file-store with
originalBlobUrl for comparison. Enhanced dropzone with multi-file count
badge and file list display.
Adds Zustand file store, useToolProcessor hook for upload/process/download
flow, and 7 settings components: resize (with social media presets), crop
(with aspect ratio presets), rotate/flip, convert, compress (quality +
target size modes), strip-metadata, and color adjustments (brightness,
contrast, saturation, channels, effects). Updates tool-page to render the
appropriate settings panel based on toolId.
Seven tool route files using the createToolRoute factory, registering 10 API
endpoints total (color adjustments covers 4 tool IDs). Also fixes the tool
factory generic to properly infer Zod output types and clamps the compress
binary-search quality to 1-100.
- createToolRoute<T> factory: handles multipart parsing, file validation,
Zod settings validation, workspace management, and error handling
- Each tool only needs to provide toolId, settingsSchema, and process function
- Registers POST /api/v1/tools/:toolId routes automatically
- Tool registry stub at routes/tools/index.ts ready for tool implementations
- Catches Sharp errors and returns clean API error envelopes (422)
- Register @fastify/multipart plugin with size limits from env config
- Workspace manager: createWorkspace, getWorkspacePath, cleanupWorkspace
- File validation: magic byte detection, format check, megapixel limit
- POST /api/v1/upload: multipart upload with validation, returns jobId + file metadata
- GET /api/v1/download/:jobId/:filename: serve files with Content-Disposition
- Path traversal guards on all file-serving endpoints
- Add @stirling-image/image-engine and sharp as API dependencies
- Add apiUpload, getDownloadUrl, apiDownloadBlob to web client
Implement custom auth using crypto.scrypt password hashing and UUID
session tokens. Includes login/logout/session endpoints, preHandler
middleware that skips public routes, and automatic default admin
creation on first startup.
- Lint & typecheck on every push/PR
- Build verification
- Docker build test
- Multi-arch Docker publish to GHCR on main (amd64 + arm64)
- Build cache via GitHub Actions cache
- Concurrency control to cancel stale runs
- Switch from Node.js ONNX to production Python ML libraries:
rembg, Real-ESRGAN, LaMa, PaddleOCR, MediaPipe
- AI models baked into Docker image for instant reliability
- Multi-architecture support (linux/amd64 + linux/arm64)
- No Docker size limits - reliability over size (expect 3-5GB)
- Object Eraser upgraded from experimental to production (LaMa)
- All AI tools now use state-of-the-art models
Critical: Use Real-ESRGAN ONNX (not UpscalerJS/TF.js) for upscaling,
mark Object Eraser as experimental, switch to Debian slim (not Alpine)
for HEIC/libraw/libjxl support, add libraw for RAW formats.
Significant: Revise Docker target to 700MB-1GB, add SSE progress events,
fix StorageAdapter to use streams, add API error schema, define SQLite
write concurrency model, add CORS/SSRF/rate-limiting specs.
Comprehensive PRD covering tech stack (Vite+React+Fastify+Sharp+SQLite),
37+ image tools across 9 categories, Docker deployment on port 1349,
UI/UX architecture, API spec, and development roadmap.
Siddharth Kumar Sah