How to Make a Professional Headshot Without a Studio

A good headshot is not a fashion shoot. It is a small, honest piece of evidence that a real person exists, does work, and can be reached. That is why the studio version matters less than people think: what actually carries a headshot is a clean source photo, controlled light, ordinary clothing, and a face that still looks like the person on their ID. You can get most of the way there with a phone, a window, and a browser tab. This guide walks through that workflow end to end, including the part most people skip, which is checking that the generated face is still recognizably yours.

The tool we use here is the AI headshot generator on wan2-6.net. It runs in the browser, you need an account and credits to generate, and the exact credit cost is shown in the studio before you submit anything, so you can decide after you see the settings rather than before.

What makes a headshot believable

Believability comes from a short list of boring things: eyes that point in the same direction as the head, a jawline that matches the neck, hair that behaves like hair rather than paint, light that falls from one identifiable direction, and clothing that fits a normal human body. Glamour is a different axis entirely. A dramatic rim light and a wind machine can make a striking image that nobody would use on a company page, because it reads as a stock photo or an actor's reel. For business use, the target is closer to "a competent colleague photographed well."

The practical consequence is that you should optimize for plausibility, not polish. Skin with visible pores and a few small asymmetries is more convincing than airbrushed skin. A slightly imperfect collar beats a collar that dissolves into the neck. When you review results, ask whether the image could have been taken by a colleague with a decent camera in a bright room. If the answer is no, the image is probably over-styled.

The source photo matters more than the prompt

No prompt rescues a bad source. The generator's job is to reinterpret lighting, framing, and clothing around the face you give it, and it can only work with what's visible. A source photo shot from below in a dim room, at 0.5x zoom, with hair covering one eye, gives the model very little to anchor on.

Shoot the source deliberately:

• Use the rear camera if you have someone to hold the phone, or the front camera at arm's length with a timer. Rear cameras are usually sharper.
• Stand facing a window during daylight. Turn off overhead lights so you have one dominant direction.
• Frame from mid-chest to just above the head, eyes on the upper third of the frame.
• Keep your expression neutral-to-slightly-pleasant. A full grin is hard to preserve; a small closed-mouth smile survives generation better.
• Take ten frames with tiny changes: chin slightly down, slightly up, head turned a few degrees. Pick the sharpest one with both eyes clearly open.
• Do not apply beauty filters, and do not shoot in portrait mode with heavy background blur. Both remove information the generator needs.

Write the prompt as a short brief, not a poem. For example:

Headshot of a person in their thirties, neutral expression with a slight smile,
facing camera, soft window light from the left, plain light-grey wall behind,
navy crew-neck sweater, shoulders square to camera, sharp focus on eyes,
natural skin texture, no heavy retouching

That prompt describes the situation you actually want. It does not describe a celebrity, a lens, or a mood. Specificity about light direction and background does more work than adjectives about quality.

A simple lighting setup you already have

One window and one wall is enough. Position yourself about a meter from the window, at a slight angle so the light hits one side of your face more than the other. That gives a gentle gradient across the cheek and a small shadow under the jaw, which is what makes a face read as three-dimensional. Face a plain wall roughly two meters behind you, or hang a bedsheet if the wall is busy.

Avoid three common mistakes. First, do not stand directly in front of the window with your back to it; that produces a silhouette. Second, do not mix window light with a warm ceiling bulb, because the two color temperatures will fight and the generator will average them into something muddy. Third, do not use a ring light at close range; it flattens the face and leaves a telltale circular catchlight that looks like a webcam.

If you shoot in the evening, a single floor lamp bounced off a white wall works as a substitute. Keep it to one light source. In the generator, describe that direction in the prompt ("soft light from the left") so the model does not invent a second source.

Clothing and background choices

Clothing should be plain, fitted, and free of logos, text, and busy patterns. Solid mid-tone colors, navy, charcoal, olive, muted burgundy, photograph reliably and read as professional without looking like a uniform. Collars help frame the face; a crew neck is fine if it sits cleanly on the shoulders. Avoid pure white, which blows out against bright backgrounds, and avoid all-black in low light, which loses the shoulder line.

For background, choose one context and stay with it. A plain wall, a softly out-of-focus office, or a neutral outdoor backdrop all work. What matters is that the background does not compete with the face for attention and does not contain recognizable text, other people, or brand marks. If you are generating multiple headshots for a team, pick one background description and reuse it verbatim.

Checking the eyes, jawline and hairline

This is the step that separates a usable headshot from an embarrassing one. Generation can drift, and the drift is usually concentrated in three places. Open each result at full size and check them in order.

Eyes: both pupils should point the same way, the catchlights should sit in the same relative position on each eye, and the eyelids should look like they belong to the same face. Asymmetric gaze is the most common failure and the most noticeable to viewers.

Jawline and neck: the jaw should connect to the neck at a plausible angle, with the shadow under the chin following the light direction you asked for. Watch for a jaw that is too sharp, too narrow, or that seems to belong to a younger face than the rest of the image.

Hairline: the hair should start where your hair starts. Receding hairlines are frequently "corrected" by models, and a filled-in hairline is an instant tell for anyone who knows you. Ears should also be plausible, since they are a common site of smearing.

A short review checklist:

1. Do both eyes point the same direction?
2. Is the hairline in the right place and the right shape?
3. Does the jaw match the neck and the ears?
4. Is the skin texture natural, with pores rather than plastic?
5. Does the light come from the direction described?
6. Is the clothing free of invented logos or garbled text?
7. Would a colleague recognize this person in two seconds?

If a result fails two or more checks, discard it rather than editing around it. Regenerating with a cleaner source or a tighter prompt is faster than patching. For a deeper treatment of what these tools can and cannot preserve, see the sibling guide on identity-safe AI photo editing.

One caveat worth stating plainly: no current system guarantees identity preservation. Every generation is an interpretation, and a small percentage of outputs will look subtly off even from a good source. That is why the review step is not optional, and why you should keep the approved original photo on file.

Making a team look consistent

Consistency across a team page comes from shared constraints, not from generating everyone at once. Fix four variables and let the faces vary: background description, light direction, framing, and clothing palette. Ask each person for their own source photo following the same shooting instructions, then run each through the generator with the same prompt template, changing only the clothing color and a short physical description.

A useful template:

Headshot of [description], neutral expression with a slight smile, facing camera,
soft window light from the left, plain light-grey wall behind, [color] crew-neck
sweater, shoulders square to camera, sharp focus on eyes, natural skin texture

Keep a shared folder of approved source photos and approved outputs, and record the prompt used for each. When a new hire joins, reuse the template rather than reinventing it. The result will not be pixel-identical across people, and it should not be, but the images will feel like they came from the same afternoon in the same room, which is exactly the impression a team page should give.

Set expectations about turnaround too. Reviewing a batch properly takes longer than generating it, so budget fifteen minutes per person for the checklist above. If you are producing headshots at any volume, check what the studio shows for credit usage before you start a batch, since the cost per submission is displayed in the interface rather than fixed in advance.

What to disclose when you publish

Generated headshots are increasingly normal, but the norms are still forming, and the safe move is a light touch of disclosure. If the image accompanies a real person and represents them, a short note in the page footer or an internal style guide entry is usually sufficient. If the image is used in a context where viewers might reasonably assume it is an unedited photograph, for example a press page or a speaker bio, say so explicitly.

Two rules of thumb. First, never present a generated headshot as documentary evidence of a specific event. Second, never generate a headshot of someone who has not consented, including composites that resemble a real individual. Beyond ethics, several jurisdictions treat synthetic depictions of identifiable people as a regulated category, so keep a signed or written approval from each person whose face appears.

Save the original source photo and the final output together, note the date and the prompt, and keep them for as long as the image is in use. If someone asks how the image was made, you will have a straight answer ready, which is a better position than reconstructing it later. If you are planning to use the workspace regularly for a team, the plans on the pricing page show what is included, and the studio itself is the place to confirm the exact credit cost of each submission before you commit.