It's worth it mainly for opticians who sell frames online or whose customers browse the website before visiting. Try-on apps for Shopify start at about $49 to $59 a month, but entry plans cover only 10 to 15 frames, so a full range costs more. For an in-store-only practice, a good dispenser usually beats AI frame advice.
The cost people miss is getting frames into the tool. Virtual try-on needs a 3D model of every frame you want customers to try. Major brands' frames may already exist in a vendor's database; your own label, independent brands and older stock usually don't, and each one needs digitising. That per-frame cost, plus the plan caps on how many frames can be live at once, decides the economics more than the monthly fee does.
What these tools actually do
- Live and photo try-on. The customer's webcam or phone camera shows them wearing a 3D frame in real time, or they upload a photo. This is the core feature.
- Face-shape analysis and frame suggestions. The tool estimates face shape and suggests frame styles. Fittingbox lists a face-shape analysis module among its website options.
- Pupillary distance (PD) measurement. Some plans estimate PD from the camera. Auglio includes "Auto-PD measurement" from its Basic plan; Fittingbox offers a PD tool on its advanced website offer but not in its Shopify app.
- Lens simulation and frame removal. Premium modules show tints and coatings, or digitally remove the customer's current glasses so they can see themselves in a new pair.
Separately, "AI frame advice" can simply mean staff using a chat assistant to suggest frames from your catalogue based on what a customer says they want. That costs almost nothing and is covered below.
The cost that decides it: plans and digitising frames
Here's what two of the main Shopify apps list on the Shopify App Store as of September 2026:
| App and plan | Monthly price | Frames live at once | Unique users a month | Notes |
|---|---|---|---|---|
| Fittingbox Bronze | $59 ($590 a year) | Up to 10 | Up to 500 | 14-day trial |
| Fittingbox Silver | $99 ($990 a year) | Up to 50 | Up to 1,500 | 14-day trial |
| Fittingbox Gold | $199 ($1,990 a year) | Up to 200 | Up to 3,500 | 14-day trial |
| Auglio Starter | $49 ($492 a year) | Up to 15 | 200 | 14-day trial |
| Auglio Basic | $119 ($1,188 a year) | Up to 100 | 1,500 | Auto-PD measurement |
| Auglio Professional | $369 ($3,708 a year) | Up to 400 | 5,000 | Auto-PD, usage statistics |
Then the frames. Fittingbox syncs automatically with its own digital frame database and says database subscribers get free digitisation of frames from the top-selling international brands. For other frames it quotes a per-frame price that runs from roughly $10 for a basic model generated by AI from two photos (ready in minutes) up to roughly $300 for premium studio modelling, which takes around three weeks after it receives the physical frames. Auglio includes basic digitisation from images in its plans and lists full 3D models from $35 per product.
A practice with 60 own-label frames that each need a full model is therefore looking at a set-up cost of a couple of thousand dollars before the first customer tries anything on, and a recurring cost each season as new frames arrive. Frame turnover matters as much as range size.
Run the sum on your own seasons. Say a practice launches two own-label collections a year of 15 frames each. At Auglio's $35 per full model, that's $1,050 a year in digitising before any plan fee. The plan cap bites too: with 80 frames already live on a 100-frame plan, 30 new frames means either retiring 10 older frames from try-on or moving up a tier, which on Auglio means the $369-a-month Professional plan. Decide in advance which frames lose try-on when new ones arrive, usually the slowest sellers, so the choice isn't made in a rush on launch day.
When it pays and when it doesn't
| Factor | Points to worth it | Points to not yet |
|---|---|---|
| Where frames are sold | A real share sold online, or reserved online for collection | All sales in the practice, after an eye test |
| Website traffic | Hundreds of visits a month to frame pages | A few dozen visits, mostly to the booking page |
| Range covered by vendor databases | Mostly major brands already digitised | Mostly own label or small independent brands |
| Frame turnover | A stable core range | Frequent new collections, each needing digitising |
| Customers | Browse on phones before visiting; buy for themselves online | Prefer to try frames with a dispenser's help |
An in-store practice, an online seller and a fashion boutique
An in-store-only independent with 600 frames on the wall. Most customers come for an eye test and choose frames with a dispenser afterwards; the website is for booking. Only a small share of the range would be covered by a vendor database, and digitising the rest would cost thousands. Verdict: skip try-on. Put the money into faster frame selection in store and better recall reminders; the tutorial on AI for recalls and bookings is a better first step.
An independent with a Shopify store selling about 80 frames online. Customers buy online with a prescription or reserve frames to collect. Online returns are mostly "didn't suit me". Verdict: worth a trial, because try-on attacks exactly the reason for returns. The worked cost is below.
A fashion-led optical boutique with strong social media. Plenty of footfall, 40 "hero" frames it promotes online, and customers who arrive having already picked a favourite. Verdict: maybe, at a small plan size. Try-on on the 40 hero frames helps turn social media interest into bookings; the rest of the range stays in store with a dispenser. The sum is modest: Fittingbox Silver's 50-frame cap holds the 40 hero frames with room for 10 new ones, at $990 a year on annual billing plus digitising for any that aren't in the vendor's database. At the same $110 of margin per frame used below, that's about nine extra frame sales a year to cover the plan.
Worked cost for the Shopify optician
Here are illustrative figures for the second practice. It sells 80 frames online: 50 from brands already in the vendor's database and 30 own-label frames that need full models.
| Item | Option 1: Auglio Basic | Option 2: Fittingbox Gold |
|---|---|---|
| Plan (80 frames needs a 100- or 200-frame tier) | $1,188 a year | $1,990 a year |
| Digitising 30 own-label frames | 30 x $35 = $1,050 | Per-frame quote; database brands free |
| Staff time: set-up and testing (about 10 hours at $25) | $250 | $250 |
| Year one total | About $2,490 | About $2,240 plus digitising quote |
To break even, the try-on needs to produce extra profit of about $2,500 in the year. If a frame sale online earns the practice, say, $110 of gross margin, that's around 23 extra sales, or fewer if it also cuts returns, each of which costs postage, handling and a lost sale. For a store selling several hundred frames online a year, that's a realistic test to set. For one selling 50, it isn't.
The in-store case: customers who can't see without their glasses
There is one in-store situation where try-on earns its place even in a practice that sells nothing online. A customer with a strong prescription takes off their glasses to try a frame with demonstration lenses, looks in the mirror, and sees a blur. They choose frames almost blind, or rely on a partner's opinion, and some of them come back unhappy when the finished glasses arrive.
A tablet or screen running try-on with a frame-removal feature changes that: the customer keeps their own glasses on, the software removes them from the image, and they see a sharp picture of themselves in the new frame. Fittingbox markets an in-store version as well as its website tools; ask any vendor for a demonstration with a customer wearing strong lenses, because frame removal is harder to do convincingly than basic try-on. For a practice with many high-prescription customers, a handful of remakes avoided each year may justify a small plan on its own, even if the website never uses it. Ask the vendor how in-store pricing differs from the web plans, because it is often quoted separately.
AI frame advice from a chat assistant: a sample and what to fix
The cheapest form of AI frame advice is a staff tool, not a customer-facing one: a chat assistant given your current frame list, asked to shortlist frames for a customer. An illustrative prompt:
Here is our current frame list with code, shape, material, colour,
size (lens width-bridge-temple) and price: [paste list].
A customer says: "I want something lighter than my heavy black frames,
not too round, maybe a warmer colour. Budget up to about $180. I wear
them all day for computer work." Their current frame size is 52-18-140.
Shortlist 5 frames from our list only. For each, give the code, one
line on why it fits what they said, and flag any that are more than
2 mm different in lens width from their current frame.An illustrative response (extract):
1. TR-214 Tortoise, rectangular acetate, 52-17-140, $165. Warm colour,
lighter than a heavy black acetate, similar size.
2. TI-088 Honey, titanium panto, 51-19-145, $179. Very light for all-day
wear; soft shape without being fully round.
3. ... flatters a square jawline ...What to fix: the third suggestion slipped into face-shape flattery the customer didn't ask for, which is exactly the kind of generic advice that puts people off. Tell the assistant to use only what the customer said. Check sizes against your stock system, since the list you pasted may be out of date. That one shows up quickly: a list pasted in spring will happily shortlist a frame discontinued in summer, and the customer asks to try on something you no longer have. Export the list fresh each week with a stock column, and tell the assistant to skip anything showing zero. And keep the final choice with the dispenser, who can see how a frame actually sits. The broader list of everyday uses is in eight practical AI uses for an independent optician.
Measurements, privacy and the dispensing line
Camera-based PD measurement is convenient for online orders of single-vision glasses, but treat it as an estimate. Fitting heights for varifocals and bifocals, and fine adjustments for strong prescriptions, still need a proper in-person measurement. If you sell online, say clearly which lens types you'll accept a camera PD for, and which need a visit.
Illustrative wording for the product page, adapted to your own rules: "You can use our camera measurement for single-vision glasses. For varifocals, bifocals or a prescription stronger than [your threshold], we'll need to measure you in the practice: book a free 10-minute fitting here." Put the same rule into checkout, not just the FAQ, so an order for varifocals with a camera PD is held for a phone call rather than sent straight to the lab.
Privacy needs attention because try-on processes images of customers' faces. Fittingbox states that its try-on runs in the user's browser, doesn't store the image, and doesn't collect facial geometry or other biometric data. Not every vendor says the same, and try-on tools have attracted biometric-privacy class actions against retailers in some jurisdictions. Read the vendor's privacy terms (the guide on what to check in an AI tool's privacy policy helps), update your own privacy notice to mention try-on, and ask the vendor in writing whether any face image or measurement ever leaves the customer's device.
The question to send, word for word: "Does any image, video frame or facial measurement leave the customer's device? If so, what is sent, where is it stored, for how long, and who can access it?" Replies vary in usefulness. An illustrative vague one: "We take privacy seriously and comply with all applicable regulations." That tells you nothing. An illustrative useful one: "Try-on runs in the browser. No image or facial geometry is sent to our servers. The PD value is passed to your store only if the customer adds it to an order." If you get the first kind, ask again, and don't go live until you get the second.
A 30-day test with numbers to track
Both apps above offer 14-day trials, which is short for a website test. Negotiate a longer trial, or pay for one month on the smallest plan that fits your top frames.
Before any customer sees it, check every digitised frame against the real one. Put the physical frame on a colleague, open try-on on the same colleague, and compare the width at the temples, where the bridge sits and the colour. In one illustrative check, 4 of 30 own-label models built from two photos came out noticeably flatter in colour than the real tortoiseshell, and one rendered about a size wider than the frame, exactly the gap that produces a "looked different online" return. Those five were re-photographed or sent for a fuller model; the rest went live. Then track:
- Try-on sessions and the share of product-page visitors who use it. Under 5% suggests customers don't want it or can't find it.
- Add-to-basket rate for frames with try-on against similar frames without it over the same weeks.
- Reservations or bookings that mention a frame the customer tried on.
- Returns for "didn't suit me" on try-on frames versus the rest.
- Staff time spent on set-up, frame updates and customer questions about the tool.
An illustrative month for the Shopify practice, with try-on on its 30 best-selling frames, compared with 30 similar frames without it:
| Measure | Frames with try-on | Similar frames without |
|---|---|---|
| Product-page visits | 1,240 | 1,180 |
| Try-on sessions | 104 (8% of visits) | n/a |
| Add-to-basket rate | 4.0% | 2.7% |
| Orders | 21 | 15 |
| "Didn't suit me" returns | 2 | 5 |
| Staff time on the tool | 6 hours, mostly set-up | n/a |
Encouraging, but small. Six extra orders and three fewer returns in one month could be chance, so the honest verdict is "extend the test by a month and add 20 frames", not "roll it out to everything". If the second month points the same way, the break-even of about 23 extra sales a year from the worked cost above looks reachable, since six a month would be over 70 a year.
If the try-on frames sell measurably better and returns fall, extend it to more of the range and budget for digitising new collections. If sessions are low and sales unchanged, cancel, and try AI model photos on product pages instead, a cheaper route covered in AI model shots without a photoshoot.
What opticians ask before adding try-on
Can virtual try-on work on the practice's own website if it isn't on Shopify?
Yes, but usually through a website integration rather than an app store listing, and often on a quoted rather than published price. Fittingbox, for example, offers website integrations for other platforms alongside its Shopify app. Ask about set-up work, which frames are included, and whether you or the vendor maintains the integration when your site changes.
Does try-on reduce returns for online glasses sales?
It can, if returns are mostly about customers not liking how frames look on them. It won't help with returns caused by fit, comfort or lens problems. Measure your return reasons before and after rather than assuming, and compare frames with try-on against similar frames without it over the same period.
Is a face-shape frame recommendation accurate?
Treat it as a style convention, not a measurement. Rules like 'round frames for square faces' are fashion guidance, and many customers like frames the rules say they shouldn't. Face-shape tools are useful as conversation starters and for customers who feel lost among hundreds of frames, but a dispenser's eye and the customer's own taste should win.
Further reads
- Shopify Magic vs ChatGPT for Product Descriptions — Writing frame product descriptions with Shopify Magic or ChatGPT.
- Explaining Lens Options and Eye Test Results With AI Help — Using AI to explain lens options and test results to customers.
- How to Check AI-Generated Images for Errors Before Posting — Checking AI-generated product images before they go live.
- How to Get Your Products Recommended in ChatGPT Shopping — Getting your frames recommended when customers shop through ChatGPT.
- How to Review an AI Tool After 90 Days: Keep, Fix or Cancel — How to decide whether to keep a tool after the trial.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Fittingbox Virtual Try-On listing on the Shopify App Store, Fittingbox 3D digitisation and offer comparison pages, and Fittingbox's data-privacy statement on try-on; Auglio Eyewear Virtual Try-On listing on the Shopify App Store; a law firm alert on biometric privacy class actions against virtual try-on retailers (all checked 27 September 2026).