One scene photo plus one identity portrait
Use the standard Photo Face Swap workspace. Put the scene in Source Photo and the identity portrait in Target Face.
Deep face swap photo guide
How to deep face swap two photos online, step by step
A deep face swap uses two separate images with different jobs: Source Photo is the scene you keep, while Target Face is the identity you apply. Confirm both previews, run one deliberate test, then diagnose pose, lighting, hair, occlusion, and edge problems one variable at a time.
Direct answer
What “deep face swap” means in this two-photo workflow
Source Photo is the main image whose scene you keep. Target Face is the separate portrait whose identity you apply. In this guide, “deep face swap” means an AI face replacement that preserves the target scene rather than a broader character or scene replacement. It is not a fifty-fifty merge: the source keeps the body, hair, pose, clothing, light, framing, and background, while the target-face image supplies facial identity cues.
Before you generate: check the labels above both previews. If the scene image appears under Target Face, or the identity portrait appears under Source Photo, swap them back.
Five-step photo workflow
Place the scene in Source Photo and identity in Target Face
Do not treat the two uploads as interchangeable. Source Photo is the scene you want to edit. Target Face is a separate portrait whose visible facial landmarks guide the replacement.
- Confirm permission and choose Source Photo. Use a photo you are allowed to edit. Prefer a scene where the face is large enough to inspect and important landmarks are not hidden.
- Choose one clear Target Face. Use a sharp, unobstructed identity portrait with a pose and expression reasonably close to Source Photo. Avoid group shots, heavy filters, and tiny faces.
- Upload both photos to the labeled slots. Place the scene image in Source Photo and the identity portrait in Target Face. Confirm both previews before generation.
- Review cost and the trial image watermark. A photo face swap currently costs 3 credits. Before any completed credit purchase, the trial image visibly displays DeepSwapAI.com and the account's 100% complete email address. Every character is visible; the address is not masked, shortened, hashed, or omitted. A trial image cannot be delivered without that complete-email watermark. After any completed one-time credit purchase, future image exports are watermark-free; existing trial files do not change.
- Generate, inspect, and change one input variable. Review identity, face edges, light, hair, occlusion, and scene continuity. If you retry, change one Source Photo or Target Face variable so you can identify what improved or weakened the result.
Image-only export rule: the complete-email statement above applies to trial image exports and does not describe GIF or video watermark behavior. Read the trial image watermark and paid-unlock guide before generating or sharing.
Input priority
Fix the highest-impact conflicts first
Start with identity visibility and pose. Fine texture cannot compensate for a face that is too small, blurred, hidden, or facing a very different direction.
| Variable | Prefer | Avoid | Why it matters |
|---|---|---|---|
| Identity reference | One sharp, well-lit, unobstructed face | Group shots, heavy filters, tiny faces | Clear landmarks reduce identity ambiguity. |
| Pose | A similar yaw and head tilt to the target | Front-facing reference for an extreme profile | Large angle conflicts can distort facial structure. |
| Expression | A neutral or target-like expression | Closed eyes or exaggerated expression unless required | Mouth and eye geometry affect the perceived match. |
| Lighting | Even light with visible facial detail | Crushed shadows, blown highlights, colored face light | The target scene still determines light and skin shading. |
| Occlusion | Eyes, nose, mouth, and jaw mostly visible | Hands, hair, glasses glare, masks, or objects covering landmarks | Hidden landmarks leave fewer reliable identity cues. |
| Resolution | A face with enough pixels to inspect at normal size | Upscaled thumbnails or compression artifacts | Upscaling does not recreate missing identity detail. |
Fastest useful test: use one Target Face that matches the Source Photo pose and one that is more frontal. Compare the eyes, mouth, jawline, and scene continuity rather than judging only overall resemblance.
Choose the right workflow
Two separate photos and two people in one photo are different tasks
Two or more people already in one group photo
Use Multiple Face Swap. It detects people in the group image and maps each selected face to a separate reference.
The same face across several separate photos
Use Batch Face Swap and review the batch versus multiple-face decision guide before uploading a set.
Diagnosis
Read the symptom, then change one variable
The identity looks weak
Use a closer, sharper reference with fewer filters. Make sure the face occupies enough of the image to show eye, nose, mouth, and jaw detail.
The face shape looks unstable
Reduce the angle difference between reference and target. Extreme profiles and steep head tilts are harder than near-matching poses.
Skin tone or lighting looks pasted on
Choose a target with more even facial light or a reference without strong colored lighting. Preserve the target scene as the lighting source of truth.
Hair or glasses look wrong
Hair and accessories belong to the target scene. Select a target where they do not hide critical landmarks, and inspect edge transitions at normal viewing size.
Expression feels unnatural
Use a reference with a compatible mouth and eye state. If the target expression is extreme, test a closer expression before changing other variables.
The result changes across frames
Move to the video guide or GIF guide and inspect motion blur, turns, occlusion, lighting changes, and loop seams frame by frame.
Review checklist
Accept the result only after four passes
- Identity: compare eyes, brows, nose, mouth, jaw, and age cues at a normal viewing size.
- Scene continuity: check hair, ears, glasses, hands, shadows, and background edges.
- Technical quality: inspect blur, compression, resolution, and any abrupt texture boundary.
- Permission and disclosure: confirm you have consent and use the privacy checklist before sharing a result that could mislead viewers.
For one image, open the photo face swap workspace. For a set, test one representative target before using the batch workflow.
Two-photo questions
Answers before you upload
Which image goes in Source Photo?
Source Photo is the main scene image you want to edit. Its body, hair, pose, clothing, lighting, framing, and background provide the visual context for the result.
Which image goes in Target Face?
Target Face is the separate identity-reference portrait. Use one clear, unobstructed face with an angle and expression reasonably close to the person in Source Photo.
Does a two-photo face swap merge both photos equally?
No. The uploads have different roles: Source Photo supplies the scene, while Target Face supplies facial identity cues. Reversing them changes the intended task.
Can I swap two people who are already in the same photo?
Use Multiple Face Swap for a group photo. It detects the people in one image and lets you map a separate identity reference to each selected face. The standard photo workflow is for one scene image plus one identity reference.
How much does one two-photo face swap cost?
The current photo workflow costs 3 credits per generated result. The live workspace shows the final credit amount before submission.
Record the review
Use one rubric instead of changing the standard after every result
The free face swap quality scorecard converts the visible checks above into seven documented criteria for photos, videos, and GIFs. It calculates locally, exports JSON or CSV, and keeps consent, identity mapping, and disclosure as critical gates that a numerical score cannot override.
How this guide was made
Product workflow evidence, not a synthetic benchmark
The DeepSwapAI Product Team reviewed the live Source Photo and Target Face labels, upload order, 3-credit cost, trial image export rule, and observable quality variables, then connected those facts to the batch, group-photo, video, GIF, and output-scorecard workflows. This guide does not promise that one input will produce a particular result or claim that “deep” means a specific private model architecture. The linked scorecard is a transparent human-review rubric, not a biometric metric or provider leaderboard. Product facts and limits were reviewed on August 13, 2026; see the verification methodology.
Run one deliberate two-photo test
Place the scene in Source Photo and a compatible identity portrait in Target Face, confirm the 3-credit cost and complete-email trial watermark, then inspect the result before scaling.