What Smart Lock Is Best for Glass Doors?
A customer asks this question every week, and the honest answer starts by refusing the question. There is no single best glass-door smart lock. There is only the lock that survives your door.
Short answer: the best smart lock for a glass door is a purpose-built glass-door lock whose published fit envelope — supported glass thickness, clamp geometry, edge clearance — covers your measured panel, matched to your door's construction type (framed or frameless, single or double-leaf, hinged or patched) and to the credential your users can actually operate daily. Retrofit tempered glass almost always forces a clamp-on design, because glass drilled after tempering shatters.
That answer is a sequence, not a product name. Buy a wooden-door lock and the hard part is choosing features. Buy a glass-door lock and the hard part is discovering whether your door even permits one — before the shipment arrives, not after. This document walks the full specification sequence used by buyers who get it right the first time: seven stages, each with an entry condition and an exit gate. Work them in order. Skipping ahead is where the money gets burned.
The Whole Sequence on One Page
Here is the method end to end. Each box is a stage gate: you exit one stage only when its output document exists and its check has passed.
flowchart TD
S1[Stage 1 — Door census: framed vs frameless, single vs double, hinge vs patch fitting] --> S2[Stage 2 — Mount architecture: clamp-on vs pre-drilled mount]
S2 --> S3[Stage 3 — Credential decision: fingerprint / passcode / RFID card / app-Bluetooth / key override]
S3 --> S4[Stage 4 — Fit-envelope verification with the factory: measured glass thickness vs published band]
S4 --> S5[Stage 5 — Sample order and bench verification]
S5 --> S6[Stage 6 — Install sequence on site]
S6 --> S7[Stage 7 — Post-install check and handover]
S4 -->|measured value outside band| X[Stop: re-scope door or re-specify hardware — do not order]
S5 -->|sample does not clamp true| X
Two exits lead to a hard stop, and that is the point. On a wooden door, most specification errors are correctable on site with different screws or a different strike. On tempered glass, an incorrect fit dimension is not correctable at all — the panel cannot be re-drilled, re-thickened or re-tempered to suit hardware you already paid for. The stages before order placement carry the entire risk.
Stage 1 — Census the Door Before You Census the Catalog
Entry condition: you have physical or documented access to the actual doors. Exit gate: every opening is logged with construction type, leaf count, top-and-bottom hardware, and glass condition.
The glass-door category exists because glass doors are not one product. They are at least four build-ups, and each one changes which lock architectures are eligible:
| Door construction | What you see on site | What it forces in the lock | Common trap |
|---|---|---|---|
| Frameless, hinged with patch fittings | Bare tempered panel, square stainless patches at top/bottom carrying hinges and a floor spring | Clamp-on lock body positioned away from patch zones; double-side access through the glass | Ordering a lock whose body overlaps the patch-fitting band |
| Frameless, pivot or floor-spring swing | Single pivot point, no visible frame | Same clamp logic, plus swing-side check: reader face and battery compartment must land on the correct side | Ignoring which way the leaf swings and mounting the reader on the public face |
| Framed aluminum or uPVC glass door | Glass set inside a metal stile of some width | Lock choice is driven by frame profile width, not glass — narrow-edge mortise hardware, not clamps | Buying a clamp-on glass lock for a door that actually has a frame deep enough for mortise hardware |
| Double-leaf glass doors (frameless or framed) | Two leaves meeting at center stiles; often one active, one passive | Center-stile hardware; coordination of active/passive leaves; possibly two units | Assuming one lock secures both leaves; the passive leaf needs its own coordination hardware |
A fifth variant shows up in older storefronts: glass panels set in timber frames. Treat those as framed doors and measure the stile.
Our catalog carries three dedicated glass-door SKUs — the SLK-G200, the B09 listed for double-opening face-recognition applications, and one further model — plus adjacent options that buyers sometimes route to glass-adjacent doors. That is deliberately a small shelf, because a glass-door lock is a fitting problem wearing a lock costume. The full category view is on the glass door locks page; per-model fit bands are transcribed at Stage 4, not assumed here.
If the census is documentary rather than physical — a door schedule instead of a site visit — order a site survey before anything else. The glass door smart lock site survey guide covers exactly what to measure and with what instrument.
Stage 2 — Decide the Mount Architecture: Clamp-On vs Drilled
Entry condition: door construction logged for every opening. Exit gate: a mount architecture chosen per opening, with the drilling question answered definitively.
This is the fork that separates glass-door hardware from everything else in the building.
Tempered glass is strength-balanced through its whole surface. Any hole, notch or edge grind after tempering breaks that balance and the panel disintegrates — sometimes instantly, sometimes hours later. So the drilling question is really a timing question:
- Was the hole specified before the glass was tempered? If the doors were custom-ordered with pre-cut lock holes and you hold the original fabrication drawing, drilled-mount hardware is possible. Re-verify hole position against the candidate lock's hole pattern with the drawing, not from memory.
- Is the glass already installed? Then it was tempered as-installed-dimensions, and no hole can be added. Clamp-on is the only architecture.
| Dimension | Clamp-on glass-door lock | Pre-drilled mount lock |
|---|---|---|
| Glass requirement | Bare tempered panel within the published thickness band; sound edges | Factory-cut hole to drawing; panel cannot be changed without reordering glass |
| Retrofit suitability | The standard answer for existing doors | Effectively new-build or full door replacement |
| Fixing principle | Interior and exterior bodies sandwich the panel through padded contact faces, drawn by through-bolts | Hardware passes through the pre-cut hole |
| Failure mode if wrong | Under-clamped: body rotates/sags; over-clamped on out-of-band glass: edge damage | Hole pattern mismatch: unusable, no on-site correction |
| Consequence of error | Re-torque, re-pad, or return unit | Replace the entire door |
| Who specifies it | Most commercial retrofits, most of our glass-door inquiries | New builds where architect froze door drawings early |
One clarification that stops expensive confusion: "drilled" in glass-door conversations sometimes means drilling the frame or the floor spring cover, not the glass. Small pilot holes into aluminum stiles or patch-fitting cover plates can be legitimate. Drilling into the glass panel itself after tempering is never legitimate. Whatever the marketing page says, ask the direct question in writing: does installation require any modification of the glass panel? Anything other than "no" on an existing door disqualifies the product.
Stage 3 — Choose the Credential for the People Who Will Use It
Entry condition: mount architecture fixed. Exit gate: credential set selected per door group, with an override path and an issuance workflow agreed.
Glass-door locks live in offices, retail fronts, clinic partitions, co-working suites — places with more foot traffic and less privacy expectation than bedrooms. Unlocking methods available across our catalog generally include fingerprint, passcode, RFID/IC card, mechanical key, and app control over Bluetooth on supported models (TTLock, Tuya and Doodle ecosystems appear across the range; per-model platform support and whether any given SKU offers Wi-Fi connectivity must be confirmed on that product's page). Face recognition exists on a few catalog models, including the B09 double-opening glass application noted above. Score the credentials against the actual door, not against a features list:
| Credential | Best-fit glass-door situation | Watch-outs |
|---|---|---|
| RFID/IC card | Offices and clinics with badge systems already in staff pockets | Card fatigue — staff carry one more card unless the new lock reuses the existing badge format; verify format compatibility at Stage 4 |
| Passcode | Retail and low-security partitions where visitors never get codes | Keypad exposure to shoulder-surfing on storefront glass; code churn discipline required |
| Fingerprint | Staff-only back entrances, one stable user pool | Read reliability through gloves or wet hands; enroll every regular user before go-live |
| App / Bluetooth | Distributor demos, small offices, users who resist carrying cards | Phone-battery dependency makes the mechanical override non-negotiable; confirm per-model app platform on the product page |
| Mechanical key override | Every door, as the failure backstop | On glass locks the override keyway sits in the exterior body — check it is a real cylinder, not a cosmetic cover |
The double-leaf question belongs here too. Two active leaves with independent locks can fight at the center stile; a passive-leaf coordinator (or locking only the active leaf) is usually the saner plan. For frameless doors specifically — where pivot geometry, patch-fitting clearance and swing side interact with every one of these choices — the companion article in this series, What Smart Lock Is Best for Frameless Glass Doors? (What Smart Lock Is Best for Frameless Glass Doors?), takes that branch to full depth. This article stays on the general glass-door method.
Stage 4 — Verify the Fit Envelope Against the Factory, on Paper, in Writing
Entry condition: construction type, mount architecture and credential set fixed. Exit gate: a written confirmation from the factory or supplier that your measured values fall inside the published envelope for the exact SKU — or a documented stop.
This is the gate where glass-door orders are actually won or lost. A wooden-door fit envelope is roughly one axis (door thickness plus backset). A glass-door fit envelope is at least four axes, and every axis number must come from a published specification or a factory confirmation — never from a sibling model, a forum thread, or memory:
- Supported glass thickness band — with the clamp's adjustment margin inside it. Commercial frameless panels most often fall in the 8–12 mm nominal region, with 6 mm, 15 mm and 19 mm panels appearing in storefronts, partitions and older stock. Nominal is not actual: a "10 mm" panel reads 9.8–10.2 mm in the field, and the published band may be written either way. Measure with a digital caliper on the bare glass, at multiple points, at the actual mounting height. The full measurement protocol — instruments, film interference, grounded-edge traps, what to do when a panel lands on the edge of the band — is developed in the SLK-G200 glass door thickness guide.
- Edge and corner clearance — distance from panel edge to the intended lock centerline, and whether the clamp's footprint collides with patch fittings, hinge bands or frame stiles.
- Gap and meeting-stile dimensions — especially on double leaves and on doors with floor springs where the bottom gap hides under a cover.
- Swing side and body orientation — which face gets the reader, which gets the battery compartment, and whether the unit is handed or symmetric.
For our own catalog, the fact position is explicit and unchanged: the supported thickness band, mounting style, credential set and power configuration of the SLK-G200 — and of the B09 and the third glass SKU — are Check the model’s published specification: transcribe them from each live product page before quoting or ordering, and do not port numbers across models. A buyer who reads a G200 band off a hotel-lock datasheet is the buyer who ends up with a clamp that will not close on a 12 mm panel.
What should the confirmation look like? Copy the shape of a proper first-party spec block. On the hotel side, we publish exactly that kind of block for the SLK-9936S: door thickness 36–100 mm · RFID Mifare 1 @ 13.56 MHz · read distance 0–5 cm · read time <0.25 s · power 4 × AA · battery operating cycles >15,000 · audit/event records up to 1,000 · 5-pin cylinder · 3-point latch · stainless steel + alloy body · multiple finish options · mechanical emergency access · multilingual management software · standalone or centralized database modes · positioned for 1–5 star hotels. That is a fit envelope plus operating envelope in one signed block. Demand the glass-door equivalent from any supplier — for SLK-9936S the block is quoted as-is from our product page and is not a template that transfers to glass SKUs, whose numbers live only on their own pages.
RFQ worksheet — the lines a glass-door inquiry must carry:
| # | Line item | Your value (fill in) | Source |
|---|---|---|---|
| 1 | Door construction type | framed / frameless-patched / frameless-pivot / double-leaf | Stage 1 census |
| 2 | Measured glass thickness, actual, at mount height | ___ mm (nominal: ___) | Digital caliper, ≥3 points |
| 3 | Required SKU | SLK-G200 / other glass SKU | Stage 3 decision |
| 4 | Published thickness band of that SKU | ___ – ___ mm | to verify in writing — transcribe from live product page |
| 5 | Edge clearance and patch-fitting positions | ___ mm | Site sketch |
| 6 | Badge format to reuse (if any) | e.g. existing office credential type | Facility manager |
| 7 | App/Bluetooth platform required | confirm per-model support on product page | Product page |
| 8 | Wi-Fi required? | check the model’s published specification whether Wi-Fi is offered for this model | Product page |
| 9 | Mechanical override key quantity | ___ | Policy decision |
| 10 | Written fit confirmation from factory | yes / no | Stage 4 gate |
Line 10 is the exit gate. No written confirmation, no purchase order.
Stage 5 — Order Samples and Bench-Verify Before Committing Volume
Entry condition: Stage 4 paper check passed for the exact SKU and measured values. Exit gate: a sample unit clamps a panel of your actual thickness — ideally a scrap of your actual glass — and every claimed function is exercised by the people who will use it.
Samples are the cheapest insurance in this product class. A sample order costs a freight bill; a wrong container costs the freight bill, the merchandise, the project delay, and the relationship with your own customer. Order through the sample order page with the Stage 4 worksheet attached so the factory ships the configuration you actually verified.
Run the bench check as a gate list, not an unboxing video:
- Clamp fit on real thickness. Mount the sample on a panel of your measured value (or the closest genuine panel you can source). Confirm the clamp closes with adjustment margin still available — a unit that just barely closes on the band edge is a field problem waiting for the panel that reads 0.2 mm high.
- Body position vs obstructions. Hold the assembled unit against the door at the planned height and check patch fittings, frame stiles and swing clearance with the door in motion, not at rest.
- Every credential, every user type. Enroll a fingerprint for the driest-handed staff member. Test the keypad through the glove season if your site has one. Read a card at realistic distance and angle. Pair the app on both an old Android and a recent iPhone if the deployment is mixed.
- Override key works and is counted. Turn it fully, twice, before the unit goes anywhere near a real door.
- Sound and power behavior. Note latch-motor noise against clinic or office tolerance. Ask how the unit signals low battery and what the emergency power terminal accepts.
- Instruction quality. The install sheet is a preview of your installer's experience. If page three contradicts page one, you have learned it now for the price of one sample instead of later at full volume.
A sample that fails the bench check is a success — it failed on your workbench instead of on a customer's storefront. Return to Stage 2 or Stage 4 with the failure documented.
Stage 6 — Install Sequence on Site
Entry condition: verified SKU, verified panel, install kit complete. Exit gate: unit mechanically true, credentials enrolled, first service interval logged.
Clamp-on installation is low-tool but high-discipline. The sequence below is the one we recommend writing into the job card; timings vary with door count and site access, so the table states dependencies rather than hours.
| Step | Action | Depends on | Why the order matters |
|---|---|---|---|
| 1 | Clean mounting faces; inspect edges for chips | Panel cleared by Stage 4/5 | Grease under the pads = slow creep of the body weeks later |
| 2 | Position interior body; mark through-bolt holes against published hole template | Step 1 | Template marks made before disassembly avoid a second trip to the ladder |
| 3 | Feed bolts through; seat exterior reader body with pads aligned | Step 2 | Bolt-thread cross on the first start is how aluminum threads die |
| 4 | Draw up evenly, diagonally, to the specified sequence — no more | Step 3 | Uneven draw-down tilts the body; over-torque on out-of-band glass cracks edges |
| 5 | Fit strike/keeper on frame, floor spring cover or opposing leaf | Step 4 complete | Latch alignment is set by the fixed body, never adjusted by force |
| 6 | Power up; enroll master credential; test latch and (where applicable) motor travel | Step 5 | Every function is proven while the ladder is still at the door |
| 7 | Program user credentials per Stage 3 policy; issue override keys under control | Step 6 | Keys and cards issued before Step 6 become inventory to reconcile later |
| 8 | Post-install check (below) sign-off | Step 7 | The gate that makes the warranty conversation factual instead of emotional |
On framed aluminum doors the branch is different: the hardware is a mortise narrow-edge unit and the fitting work concentrates on the metal stile — its width, existing mortise, and the drive case inside the profile. That branch has its own article, narrow edge locks for glass-frame doors, covering profile-width gates and what to check before ordering into an aluminum frame.
Stage 7 — Post-Install Check and Handover
Entry condition: installation signed off. Exit gate: the facility can operate the lock without the installer.
The 48-hour re-check is the habit that separates glass-door installations that age well from ones that generate service calls. Glass moves with temperature swings in ways frames never did; a clamp drawn up on a cold morning can read loose on a hot afternoon, or the reverse.
- Re-verify clamp: no rotational play when the handle or thumb-turn is thrown firmly.
- Re-verify latch strike: free travel, no scraping sound, no witness-mark smearing on the keeper.
- Test every credential type a user actually holds — not just the installer's enrolled card.
- Confirm low-battery indication is understood by whoever owns battery replacement, and that the emergency power terminal is documented in the handover file.
- Log serial, door number, mounting height, measured glass thickness, and the Stage 4 confirmation reference against each unit. When the third identical storefront door needs a replacement body next year, that log tells you which SKU band your glass actually sat in.
Handover is also where the operational questions finally have local answers: who re-enrolls a departed staff member, how often passcodes rotate on storefront units, what happens when a card fleet migrates formats. Those are facility policies, not lock features — but writing them down at handover is what stops a smart glass-door lock from being defeated by its own users within a quarter.
How the Sequence Fails: Three Illustrative Patterns
Consider a 14-door clinic fit-out where the architect's door schedule says "10 mm tempered, frameless, patch-fitted." The buyer reads that number as fact, skips Stage 1's physical census, and quotes 14 clamp-on units against a nominal band. Four doors turn out to be 11.2 mm film-laminated panels from a different vendor. Two units clamp; three sit in a warehouse as spares; one edge chips on draw-down. The entire loss traces to one unmeasured panel.
Consider a distributor demo order: samples pass the bench check on a pristine test panel, but the sales team skips Stage 5's "every credential, every user type" line. On the customer's actual door, thick gloves plus outdoor wet glass make the fingerprint reader a decoration. The demo fails commercially even though the hardware fit perfectly.
Consider a double-leaf storefront where Stage 3's passive-leaf question was never asked. Two active locks installed independently; on hot days the leaves press, and neither latch reaches its keeper. The fix is a coordinator decision that cost nothing at design time and costs a rebalance now.
None of these are exotic. They are the standard tax for skipping gates.
What "Best" Actually Meant
Rewind to the title. Across hundreds of glass-door inquiries, the orders that end quietly — no returns, no site surprises, no emergency freight — share one trait: the buyer let the door choose from the eligible set instead of letting a catalog page choose. The best glass-door smart lock is the SKU whose published, confirmed fit envelope contains your measured reality, whose mount architecture respects the tempering constraint, whose credentials match the people on the other side of the glass, and whose supplier answers Stage 4 in writing. If a supplier cannot produce that written envelope, the answer to "which is best" is simple: the one who can.
For the frameless-specific branch — pivot geometry, patch-fitting collision rules, swing-side decisions — the series continues in What Smart Lock Is Best for Frameless Glass Doors?. For per-model fit data on our own three glass-door SKUs, the product pages are the single authoritative source.
When your census worksheet and measured values are ready, send them through the quote request page — a fit review on paper costs nothing and moves the risk off the job site. Distributors building a glass-door line properly can pull the full category into view with the catalog request before the next RFQ round.

