Previews from the publisher
Restyle a permitted source photo as a quiet black-and-white studio half portrait while treating the face and glasses as the primary things to preserve. Monochrome Studio Portrait changes the setting, clothing treatment, light and crop around those visible anchors. Recognition remains best-effort: the final image needs comparison with the source, especially after a larger pose or coverage change.
Approve recognition before finish. Compare the face and glasses directly with the source, then judge lighting, wardrobe and grain. If the style is strong but facial geometry or frames drifted, keep it rejected and request a bounded correction instead of treating the attractive result as a successful identity-preserving edit.
/monochrome-studio-portrait Make my attached photo a calm black-and-white studio half portrait. Keep my face and glasses recognizable.Provide one clear authorized person photo as the edit target, not merely a style reference. State intended use or a ratio other than the 16:9 default. Optional same-person angles, desired negative-space side, clothing, background tone and grain strength can refine the brief. Do not infer sensitive traits from appearance.
Inspect the source at full size and record visible facial anchors and glasses construction separately. Check how much body is actually visible: a requested half-body reconstruction can introduce drift. Prefer a modest source-led pose, natural gaze/body angle and comfortable hair/shoulder margin rather than inventing unseen anatomy.
Choose one composition and write the Change/Preserve/Composition/Lighting/Constraints prompt. Repeat critical face/glasses details in the invariants. Use one large soft key roughly 30–60 degrees off camera, slightly above eye level, with shadow detail and believable catchlights. Preserve texture/age cues, keep 35–50% low-detail space and avoid unrequested retouching or brand text.
Use an available requested native image editor with the same prompt contract. If native editing is unavailable, the supported external fallback is the FAL GPT Image edit helper with Node.js 20+ and server-side FAL_KEY. Before using that route, explain that FAL and its model process the source photo and obtain authorization for the paid request; host-native editing may have separate plan limits.
For FAL, keep source, prompt, task and output paths distinct and save the English prompt. The helper accepts authorized local JPEG/PNG/WebP (up to 20 MB each), HTTPS references or image data URIs, with up to sixteen same-person inputs. Local references bind to byte hashes; remote references bind to URL strings, not fetched image bytes. Use one task per requested WebP output.
The helper exclusively claims a task path and locks processing before submission, records settings/input fingerprints and disables provider retry/fallback. Resume the saved request after an interruption; if no request ID was captured, preserve the uncertain record and inspect provider history before approving a new task. Do not remove a task to force resubmission. A stale sibling lock is removed only after confirming the old process is stopped.
FAL requests disable retrievable payload history and request account-restricted temporary CDN output, defaulting to one hour with a configurable 15-minute-to-one-day window. Download uses a short-lived token. These are requested provider-handling settings, not end-to-end encryption, no-processing or no-storage guarantees. Resume promptly before expiry and keep private photos/task records out of public repositories.
Inspect the result beside the source at full size and thumbnail scale. Check facial geometry, age/expression, hair/marks, exact glasses, eyes/catchlights, natural skin, anatomy, wardrobe, light direction, quiet background, grain and unwanted branding/text. The helper checks WebP container dimensions and saved bytes/hash; it does not judge identity, glasses, exact requested ratio or visual quality.
If a correction is needed, name one change while freezing approved properties; any new paid task requires authorization. Save non-destructively, preserve approved outputs and report actual format/dimensions. The FAL helper requires WebP; another format needs a separate explicit conversion/delivery step. Deliver the image, exact prompt, FAL task/request details when applicable and a concise QA note without credentials or provider tokens.
A sharp, well-lit image with visible face, eyes, hairline and complete glasses. An additional authorized view of the same person can clarify anchors. Keep pose changes modest and avoid reconstructing a large unseen body area; the workflow cannot guarantee exact identity from limited evidence.
That is an explicit priority, including frame/lens silhouette, rims, bridge, temples, size, tilt, fit and reflections. The prompt repeats those invariants and the result is compared visually with the source. No automatic script verifies the glasses, so an attractive studio finish is not enough to approve a mismatched frame.
No. It aims for recognizable visual consistency rather than forensic or biometric equivalence. Pose, angle, expression, hair or coverage changes increase drift. Face/glasses fidelity remains the primary visual acceptance check, and a targeted correction can still require another authorized generation.
Only for the supported fallback when native image editing is unavailable. That path needs Node.js 20+, FAL_KEY and authorization for each paid task; native tools follow host limits. FAL processes the source and temporarily handles output despite requested privacy/retention settings. Recorded tasks resume without automatic paid replacement, but uncertain submissions and expired results require review before a new task.
Select a star to share your experience.
No reviews yet. Ratings appear once readers share real experiences.
+7Use this Skill to create character & styled art.
+7Use this Skill to create poster.
+7Use this Skill to build website page.
+7Use this Skill to create article illustrations.
+7Use this Skill to build website page.
+8Use this Skill to create presentation.