Official Workflow GuideManual
Pixel Cafe Operating Guide
This guide turns the current product into a practical workflow: install the desktop app, import a shoot, run AI culling, review results, batch-retouch finalists, and export for delivery.
Scope
The repository currently includes a local desktop workstation, backend AI pipeline, cloud sync, and sharing features. This manual focuses on the user-facing workflow that belongs on the official site.
Desktop workstation
Built for photographers and retouchers who need import, analysis, and project data front and center.
Importlocal ingest
5 StageAI analysis
Liveproject metrics
- Import: bring in RAW or JPEG folders, generate previews, and build project structure locally.
- Analyze: run the full 5-stage AI pipeline and inspect screening, grouping, and ranking output.
- Data: desktop users should primarily see pending, starred, rejected, and training-progress data.
Web collaboration hub
Built for teammates and clients who mainly need cloud access, shared review, and collaboration data.
Cloudproject access
Multi-usershared culling
Sharereview links
- Collaborate: photographers, retouchers, and clients can review the same project together.
- Data: web users should primarily see shared status, online review state, collaborative results, and shortlist outcomes.
- Capability: the web experience emphasizes cloud review and collaboration rather than local ingest or heavy analysis.
1. Before You Install
- Recommended systems: macOS 12+ or Windows 10+.
- Supports RAW, JPEG, HEIC, and TIFF; RAW handling depends on local preview extraction.
2. First Launch
Open the app
Pixel Cafe starts a local workstation and a local backend service. On first launch, missing models may need to download or warm up before full AI features are available.
Check the basics
Review your default import path, export path, and whether optional cloud sync should be enabled. Core culling stays local by default.
Create a project
Use one project per assignment, such as a wedding, portrait session, or travel set. Clear project names make later sync and export much easier.
3. Import Photos
Import is where the workflow begins. The system scans folders, records metadata, and prepares previews for the next stages.
Choose a folder
Import the full folder for one shoot whenever possible. That gives burst grouping and scene analysis enough context.
Wait for indexing
Import runs asynchronously, so you can monitor progress without freezing the app. Watch counts, thumbnail generation, and failed items in the workstation.
Review exceptions
If some RAW files fail to preview, check camera format compatibility, file integrity, or local dependencies before starting a full culling pass.
4. Run The AI Pipeline
Pixel Cafe works best when each stage answers one question. That makes the output easier to trust, inspect, and override.
Stage 1: technical quality
Filter obvious blur, shake, and exposure problems first to reduce noise fast.
Stage 2: faces and expressions
Add face quality, gaze, and closed-eye checks for people-heavy shoots.
Stage 3: burst and similarity grouping
Cluster near-duplicates so you judge one burst as a set instead of frame by frame.
Stage 4: semantic understanding
Use deeper subject and scene understanding when you want richer AI reasoning around content.
Stage 5: aesthetic scoring and personal ranking
Use score signals plus your historical choices to surface the frame you are most likely to keep.
5. Review And Mark Up
- Start with the top frame from each burst, then expand only the groups that need a closer pass.
- Use stars, rejects, and ratings to capture your taste explicitly.
- Override AI decisions when needed; the workflow is built for collaboration between machine judgment and human judgment.
- For studio teams, let the lead photographer establish the first preference pass before handing off final review.
6. Batch Retouch
Pixel Cafe also supports lightweight finishing steps before delivery, especially for weddings and portraits.
Select finalists first
Run batch retouch only on starred or shortlisted photos so compute time goes where it matters.
Set conservative strength
Start subtle. For most client work, consistency and cleanliness matter more than aggressive smoothing.
Check hard examples
Sample a few difficult images first, such as backlit shots, group portraits, or high-ISO frames, before applying the same settings across the full set.
7. Live Collaboration And Sharing
- This is one of the strongest workflow advantages in Pixel Cafe: photographers, retouchers, and clients can review the same project together instead of passing folders back and forth.
- A reliable pattern is to let the lead photographer finish the first AI-assisted pass, then bring teammates into the shared shortlist for live review.
- For client-facing selection, confirm project state, cover image, and visible range before sending the link.
8. Export And Deliver
Choose an export path
Export finalists to a fresh delivery folder, or keep Lightroom / Capture One workflows intact with XMP sidecars.
Preserve structure
Keep project names, dates, and batch labels clear so later retouch or archival work stays organized.
Do a final check
Verify counts, thumbnails, and statuses before delivery so rejects do not slip through and selected photos are not missed.
9. Troubleshooting
- Slow import: check whether the source is on an external drive, NAS, or network volume.
- RAW preview failures: confirm the original file is readable and local preview dependencies are installed.
- Model download issues: inspect network and proxy settings, then fall back to direct download or manual model installation if needed.
- Ranking feels off: keep building star and reject samples so personal ranking has enough data.
- Desktop launch problems: use debug mode and review logs for model warmup, port conflicts, or WebView issues.