There's no single right way to source semi rimless glasses. Ask five suppliers and you'll get five answers — and every answer will happen to match whatever that supplier sells. The useful answer depends on three things: your order volume, whether you already have a brand to protect, and how much control you actually need over frame geometry.
I'm a sourcing manager handling eyewear OEM and ODM orders for eight years. I've personally made (and documented) 11 significant mistakes on half rim eyeglasses programs, totaling roughly $18,000 in wasted budget. Unused molds. A color mismatch that shipped on 600 pairs. A nylon cord spec that failed in the field. Now I maintain our team's pre-check list so nobody repeats them.
What follows is the framework I use before quoting any semi rimless eyeglasses project.
Three Scenarios Cover Almost Every Order
- Scenario A — Test order, under ~500 pairs. You're validating whether semi rimless frames sell in your market at all.
- Scenario B — Private label, roughly 500–5,000 pairs. You have a brand and an audience, and you need the frame to look like yours.
- Scenario C — Full custom, 5,000+ pairs. You need geometry nobody stocks, and you can commit to the volume.
Roughly 90% of the semi-rimless glasses programs I've worked on land in one of those three. The expensive mistakes happen when a brand applies Scenario C thinking to a Scenario A budget, or stays in Scenario A long past the point where custom geometry would've paid for itself.
Scenario A: Test Orders Under ~500 Pairs
At this volume, stop thinking about factories. You're buying optionality, not unit price.
The math isn't close. A new semi-rimless brow bar tool is a one-time cost, and most factories won't open a mold for less than 1,000–3,000 pieces. If you commit to a mold at 400 pairs, you've paid $2,000+ for the privilege of selling 400 frames — before you know whether the shape sells at all.
A sourcing partner adds a margin, often in the 3–8% range depending on the order (based on quotes we collected in Q4 2024; verify current rates). Compare that to a mold amortized across 400 pairs. The partner is cheaper. This is the one scenario where cutting out the middleman actually costs you money.
What most people don't realize is that a stock semi-rimless listing isn't really one product — it's a mold that's been run in dozens of color and finish combinations. You can usually get 3–5 colors inside one small order because the factory is changing a batch, not a die.
Two things I'd avoid at this stage. Don't ask for a custom lens groove or a non-standard cord on a test order — that's a setup change, and it pushes you into tooling territory fast. And don't skip incoming inspection just because the order is small. In 2019 we accepted a 300-pair test run on photos alone. Forty-one pairs had temple screws that wouldn't hold tension. Not a catastrophe at $2,100, but it was four weeks of sorting and re-shipping on a product we were still deciding whether to keep.
Scenario B: Private Label, Roughly 500–5,000 Pairs — Where the Math Flips
The standard advice is to go direct to the factory and cut out the middleman. At this volume I'd push back on that.
Most brands at 500–5,000 pairs don't need new geometry. They need a frame that looks like theirs. And that's a completely different purchase: color, finish, logo, temple spec, nose pad material, packaging. All of that can change on an existing semi-rimless mold — often one that's already been run for other brands.
Here's something vendors won't tell you: a large share of custom half rim glasses programs at this volume are running on molds that already exist. That isn't a scam. It's how the price point is even possible. What makes the frame yours is everything around the geometry.
Run the numbers on a new mold before you get excited about exclusivity. A simple stainless brow bar die, sample rounds, and the 6–8 week development window came in at a ballpark of $1,800–3,500 in 2023 depending on complexity. On 1,200 pairs that's $1.50–2.90 per pair before a single frame comes off the line. On 6,000 pairs it drops to about $0.30–0.60. If you're at 1,200 pairs, that money is better spent on three colorways and a better temple.
And if a vendor answers yes to every spec question without asking you a single one back, that's a red flag. Nobody who has actually run a semi-rimless line is that relaxed about groove depth.
Two things that will bite you at this stage
Color matching. Frame color is not one part. Brow bar, temples, and end pieces may be different materials with different finishing processes — anodizing, painting, PVD. They drift. For anything brand-critical, set a Delta E target; industry guidance puts tolerance under Delta E 2 for brand colors, with 2–4 noticeable to a trained observer (reference: Pantone color matching guidelines). We learned that on a 600-pair run in 2022 where the brow bar landed half a shade off the temples. Nobody at the factory considered it a defect. Our customer did.
Lens grooving. This is the semi-rimless problem nobody owns. The frame comes from one vendor. The lens comes from another. The groove that holds the nylon cord has to match the cord diameter and sit at the right depth and position in the lens edge. Too loose and the lens pops. Too tight and the lens edge chips. Neither vendor will call this their fault. Do a grooving trial with your actual lens supplier before you scale, and write the groove spec into the purchase order.
One regulatory note, since it's easy to miss: the grooved edge is a stress point, which is why a lot of semi-rimless programs default to polycarbonate or Trivex rather than standard CR-39. If you're shipping prescription eyewear into the US, lens impact resistance isn't a preference — it's a requirement under 21 CFR 801.410, which specifies the drop-ball test. Verify the current regulatory text before relying on it.
Scenario C: Full Custom Geometry, 5,000+ Pairs
Now you spend the money. Own the tooling, own the dimensions, own the cord system.
At 5,000+ pairs a mold is a rounding error, and the real asset is the geometry. Define brow bar width, end piece angle, cord anchor radius, and temple drop yourself — then put them in writing. Two things to negotiate that aren't unit price:
Mold ownership and a defined exclusivity window. If you pay for tooling, get it in the contract that the mold is yours and can be moved. Otherwise your exclusive shape shows up in someone else's catalog eighteen months later.
Spare parts supply. Semi-rimless frames fail at the cord more than anywhere else. If you're selling 5,000+ pairs into a market with any warranty expectation, you need spare cords — and you need the factory to keep supplying them after the frame itself is discontinued.
The most frustrating part of semi-rimless sourcing: the frame passes inspection, the lenses pass inspection, and the product still comes back. In 2021, one of our 1,800-pair runs had cords failing at the end-piece anchor within weeks of wear. You'd think a tensioned nylon cord would be a stable part. It isn't — material lot, UV exposure, and anchor radius all move the failure point. After the third return batch I was ready to kill the program. What finally helped was a written cord spec (material type, diameter, pull test at incoming QC) plus shipping 5% spare cords with every order.
If your frames are metal and sold in the EU, check nickel release requirements too. The reference point for items in prolonged skin contact is 0.5 µg/cm²/week under EN 1811 test methods. Titanium and some stainless grades sidestep the issue; monel and nickel-silver brow bars don't. Frame testing under ISO 12870 covers mechanical stability and related requirements — worth asking your factory which tests they actually run versus which ones they claim.
How to Tell Which Scenario You're Actually In
Four questions. Answer them honestly, not optimistically.
1. Can you sell 300 pairs of one shape in 90 days? If you're guessing, you're in Scenario A. Buy the catalog, learn the sell-through, come back.
2. Does your brand promise a specific look nobody currently sells? If yes and you're under 5,000 pairs, that's still Scenario B. You're buying differentiation through color, finish, and temple — not through new geometry. The custom shape can wait until the volume supports it.
3. Can you commit to 3,000+ pairs of a single SKU over 12 months? That's roughly where a mold starts paying for itself. Below it, you're subsidizing exclusivity you can't sell.
4. Who eats the defect cost? This one overrides the others. If returns land on you, you need to own the groove spec and the cord spec no matter which scenario you're in. If your distributor absorbs defects, you can stay in Scenario A longer than you think.
I've been on the wrong side of this under time pressure, too. In Q1 2024 we had 36 hours to lock a reorder before a Chinese New Year cut-off. Normally I'd run a second sample round. There was no time, so we went with the previous factory on trust and a phone call. It worked — but only because we'd already written the cord spec after the earlier failure. The lesson isn't move fast. It's that the specs are what let you move fast.
Bottom line: semi rimless glasses sourcing isn't one decision, it's three. Match the route to the volume, not to the quote. Price the mold, the groove, the cord, and the defect rate — not the FOB number. Plus whatever stage you're at, get the two specs that matter most, lens groove and cord, in writing. That's the part that keeps costing money when it's missing.
Pricing references above come from quotes collected between 2021 and Q4 2024 and are for general reference only; actual costs vary by factory, material, and order timing. Verify current rates and current regulatory text at the official sources before you commit.