Source copies, results, diagnosis, and QA evidence persist on the computer running the local service.
Restoration guide center
Choose a restoration path from the evidence in your photo
Start with what the source actually contains: visible damage, surviving tonal or facial signal, number of people, and the cost of getting identity wrong.
Local task recordsExternal AI processingQA
Before generation, the product discloses that images may be sent to the configured external service and waits for confirmation.
Comparison tasks require a preferred result while QA risk and output evidence stay visible.
Damage type
Match the dominant defect to the evidence that must survive cleanup.
| Recommended guide | Source cue | Review focus |
|---|---|---|
| Recover faded photos by rebuilding separation, not inventing detail | Weak contrast with faint tonal separation | Medium |
| Remove scratches without erasing the photograph underneath | Lines, cracks, dust, or surface marks | High across faces |
| Improve blurry old photos without redesigning unfamiliar faces | Scan softness, motion, focus loss, or tiny faces | High when signal is weak |
| Restore black-and-white photos without replacing their tonal history | Monochrome print, grain, or compressed gray range | Medium |
| Repair severe damage without pretending every detail survived | Mixed fading, stains, tears, exposure, and missing areas | High |
Photo type
Use the people count and documentary value to set a safer review boundary.
| Recommended guide | Source cue | Review focus |
|---|---|---|
| Restore family photos without rewriting family memory | Recognizable relatives, portraits, and albums | High identity sensitivity |
| Different old photos need different restoration boundaries | Groups, graduation archives, and mixed collections | Varies by face size |
Review and delivery
Understand the full evidence chain before accepting or storing a result.
| Recommended guide | Source cue | Review focus |
|---|---|---|
| Old photo restoration that remains accountable to the original | General workflow and restoration boundaries | Start here |
| Use AI restoration as a reviewed process, not an automatic truth | External AI, four-version comparison, and identity-risk QA | AI evidence boundary |
| Before and after evidence should include the restoration decision | Before-and-after evidence and realistic expectations | Compare source and result |
| A restoration workflow built for evidence, not blind enhancement | Diagnosis, four-version review, QA, and recovery | Workflow reference |
| Know exactly where restoration images and records can exist | Local records, external processing, and deletion scope | Storage boundary |
Restoration evidence
Keep the source, result, and review boundary together
Key principles
photo restoration guides
Begin with source signal
A dramatic defect can remain recoverable when outlines and tonal differences survive; a clean but tiny face can still carry high identity risk.
Choose by review cost
Family portraits and groups deserve more conservative comparison because an attractive identity change is still a failed restoration.
Keep privacy in the path
Local records and external AI processing are separate boundaries that should be understood before generation and cleanup.
FAQ
Frequently asked questions
Which guide should I read first?
Start with the dominant source problem. Use the general restoration guide when damage is mixed or the main risk is unclear.
Can one photo need several guides?
Yes. A faded group photo may need tonal recovery, small-face blur limits, and group-photo review at the same time.
Does choosing a guide start generation?
No. Guide pages are public information. Generation begins only inside Studio after a source is selected and the external-processing disclosure is confirmed.
Ready to begin
Start reviewing your own photoApply the same review standard to your own old photo.
Upload the source, choose a restoration route, and inspect likeness, composition, and quality evidence before delivery.