DEV Community: Zhuang Hua

Overlays Are Not Accessibility, and Scanners Are Not Compliance Certificates

Zhuang Hua — Sat, 31 Oct 2026 05:19:24 +0000

Disclosure: I'm the indie maker of A11yKeeper, so this is a soft promo, but the point stands without it.

If you build sites for clients, you've probably been asked: "Can we just add that accessibility widget and be done?" Short answer: no. Overlays are not accessibility. They patch the surface, can clash with screen readers, and don't fix the underlying markup, keyboard flow, or content problems.

Scanners are not certificates

Automated checks are genuinely useful, but they catch only part of WCAG. Many criteria (meaningful alt text, logical focus order, clear error messages, cognitive load) need a human to judge. So a clean scan result does not certify anyone as ADA compliant, and no automated tool can honestly say a site is "compliant". Be wary of anyone who claims otherwise.

What agencies actually need for client reports

A clear list of detected issues mapped to WCAG success criteria
Severity and where on the page each issue lives
Honest wording: "automated findings" plus a manual review checklist, not a compliance badge
Progress over time, so clients see fixes landing
Reports that carry your agency's branding, not someone else's

Where A11yKeeper fits

I built A11yKeeper as a small, honest scanning tool for agencies: white-label WCAG 2.2 scans you can hand to clients, with plain-language reports. It's a starting point for the audit conversation, not a replacement for manual testing. There's a founding agency offer at 50% off if you want to try it: https://a11ykeeper.com

Whatever tools you use, be upfront with clients about what automation can and can't tell them. Thanks for reading!

Why per-seat e-signature pricing is expensive for small teams (and an envelope-based alternative)

Zhuang Hua — Sat, 31 Oct 2026 05:18:20 +0000

I build small tools, and one thing that kept bugging me was how e-signature pricing is shaped. Most of it is per seat, which makes sense for a 500-person sales org and much less for a 4-person team that sends a handful of contracts a month.

The per-seat problem

Say you're a small agency or a freelancer collective. Maybe one person sends most documents, and others occasionally need to. With per-seat pricing you either pay for everyone who might send, or you share a login (which usually breaks the terms).

For reference, as of Oct 31, 2026, DocuSign's published annual-billing plans are:

Standard: $30 per user per month
Business Pro: $45 per user per month

For a team of 5 on Standard, that's $150/month, whether you send 2 documents or 200.

The envelope-pricing alternative

The other model is paying for what you actually send. That's what I went with when building FairSign:

About $0.30 per envelope
5 free envelopes to try it
Up to 10 signers per envelope
No seat license, so your whole team can send without extra per-user cost

If you send 20 envelopes in a month, that's roughly $6. Quiet month, tiny bill. Busy month, it scales with you.

When per-seat still wins

To be fair: if you send thousands of documents, need deep integrations, or have heavy compliance and workflow needs, a bigger platform may be worth it. Envelope pricing suits low-to-moderate volume and small teams best.

Takeaway

Before you commit to an annual per-seat plan, count how many documents you really send and how many people really send them. The math is sometimes surprising.

If you want to try the envelope model, FairSign is at https://fairsign.ai. Five free envelopes, no card games. I'd love feedback from other indie folks.

How to schedule Shopify sale prices so they actually end on time (without leaving variants discounted)

Zhuang Hua — Sat, 31 Oct 2026 02:37:24 +0000

I build a small Shopify app, and the problem I hear about most isn't "how do I start a sale." It's "why is my sale still running on Tuesday?"

Starting a sale is easy. Ending it cleanly — across every product, every variant, with the original prices restored exactly — is where things go sideways. This post covers why that happens and a checklist you can use whether or not you ever use a tool for it.

The problem: sales are easy to start, hard to end

On Shopify, a "sale price" is usually just two fields on each variant:

Price — what the customer pays
Compare-at price — the crossed-out "was" price

To run a sale, you lower price and set compare-at to the old price. To end it, you reverse that. Simple on one product. Messy on fifty products with sizes and colors.

Here's what tends to go wrong:

1. Hand edits don't scale to variants

A T-shirt with 4 sizes × 3 colors is 12 variants. Editing the product-level view and forgetting that one variant has a different price is very easy. Bulk editor helps, but it's still manual, and manual means someone has to be awake and at a laptop when the sale ends.

2. Stuck sale prices

The sale was supposed to end Sunday at midnight. Nobody remembered, or someone reverted most of the products. Now a handful of variants are still discounted days later — and you often find out from a customer, or from your margin report.

3. Lost compare-at prices

This one hurts more quietly. If you overwrite compare-at during the sale and don't record what it was before, you can't restore it accurately. Some products had a permanent compare-at (e.g., an MSRP) before your sale even started. Wipe that out and your "normal" listing looks different after the sale than it did before.

4. Time zones and "midnight"

"Ends at midnight" — whose midnight? Your store's time zone, your own, or your customers'? If you announced a time in an email, the prices should change at that time in the zone you announced.

5. Overlapping sales

A product in both the "Black Friday" collection and a "Clearance" sale. Which price wins? What gets restored when one ends but the other hasn't?

A checklist for timed Shopify sales

Use this before, during, and after any sale.

Before the sale

[ ] Export a snapshot of current price and compare-at for every variant in scope (CSV export from Products works). This is your undo button.
[ ] List the scope explicitly: which products, which collections, and whether all variants are included.
[ ] Note existing compare-at values. Decide whether the sale should overwrite them or leave them alone.
[ ] Write down start and end times with a time zone. Match your store's time zone setting (Settings → General).
[ ] Check for overlaps with other running or scheduled sales.
[ ] Decide on rounding (e.g., 20% off then round to .99) so prices don't land on odd numbers like $23.47.
[ ] Line up marketing (emails, banners, social posts) with the exact same end time.

When the sale starts

[ ] Spot-check a few products at the variant level, not just the product overview.
[ ] Check the storefront in a private window — the theme should show the crossed-out compare-at price.
[ ] Confirm the collection or discount banner points to the right products.

When the sale ends

[ ] Restore prices from your snapshot — every variant, not just the default.
[ ] Restore the original compare-at values (including ones that existed before the sale).
[ ] Remove sale banners, badges, and announcement bars.
[ ] Spot-check again: search the admin for products with a compare-at higher than price that shouldn't have one.

After the sale

[ ] Keep the snapshot file for a while in case something surfaces later.
[ ] Write down what went wrong so the next sale has fewer surprises.

Automate the boring, risky part

You can do all of the above by hand. Lots of merchants do. But the riskiest step — reverting every variant at an exact time — is also the most boring one, and that's exactly the kind of thing that's worth automating.

That's why I built SaleGuard. It lets you schedule a sale with a start and end time, applies prices across the variants you pick, saves the original price and compare-at values, and restores them when the sale ends. The goal is simple: the sale ends when you said it would, and your catalog goes back to exactly how it was.

🛍️ Shopify App Store: https://apps.shopify.com/saleguard-flash-sale-editor
🌐 Website: https://getsaleguard.com
▶️ Short demo: https://youtu.be/8PM7dKO3Lq4

No pressure — even if you never install anything, steal the checklist. The "export a snapshot first" habit alone will save you a bad Monday.

Wrapping up

Timed sales fail at the edges: the forgotten variant, the overwritten compare-at, the midnight nobody agreed on. Make a snapshot, be explicit about scope and time zone, verify at the variant level, and restore from your snapshot instead of from memory.

If you run Shopify sales and have a horror story (or a better process), I'd genuinely love to hear it in the comments — it directly shapes what I build next.

What does an e-signature actually cost? Why I built FairSign to charge at cost

Zhuang Hua — Fri, 30 Oct 2026 23:03:37 +0000

If you send a handful of contracts a month, seat-based e-signature plans can work out to several dollars per signature. The actual cost of an envelope (storage, email delivery, audit trail, PDF processing) is far lower.

What goes into one envelope

Storing the PDF and the signed copy
Sending emails to each signer
Recording an audit trail (who signed, when, from where)
Producing the final signed PDF and completion certificate

None of these needs a per-seat subscription at small volume.

How FairSign prices it

FairSign (https://fairsign.ai) charges at cost:

5 envelopes free, then about $0.30 per envelope
Up to 10 signers per envelope
No seats or bundles. Each envelope comes with a receipt that shows what it cost

Upload a PDF, place the signature fields, add signers, send. It works in the browser today.

👉 Send your first envelope free: https://fairsign.ai/documents

Comparing it with DocuSign or another tool? Tell me in the comments what you'd need to switch. It shapes what I build next.

Side note: we're also testing a Google Forms → PDF add-on. If that's useful to you, there's an optional waitlist at forms.usefairsign.com.

I built a Shopify app that restores sale prices on time (before BFCM)

Zhuang Hua — Fri, 30 Oct 2026 22:51:19 +0000

BFCM is when sale prices get stuck.

I built SaleGuard for Shopify merchants who schedule sales by the minute: it snapshots every original price and compare-at price first, changes the real product price for the sale window, then restores and checks each variant when the sale ends. Prices you edit by hand mid-sale are kept and listed in the report.

Free plan includes snapshot + restore
Starter $9.99/mo · Pro $19.99/mo (Shopify Billing)
Only 2 permissions; no customer data

Install on Shopify:
https://apps.shopify.com/saleguard-flash-sale-editor

Site / demo: https://getsaleguard.com

I'm an independent developer (not affiliated with Shopify). Feedback welcome — especially if you've been burned by a sale that didn't end cleanly.

I built a free WCAG 2.2 checker that's honest about what automated scans miss

Zhuang Hua — Wed, 28 Oct 2026 06:33:18 +0000

Agencies get asked "is our site accessible?" all the time. Most answers are either vague ("we follow WCAG") or a five-figure manual audit.

I built something in between: A11yKeeper — paste a URL, it runs axe-core across that page plus up to 9 linked pages (~1 min), groups issues by severity, and gives a plain-English fix for each.

What I tried to get right

It says clearly that a clean scan ≠ accessible. Deque's 2021 data puts automated detection at ~57% of issues by volume; every report ends with what a human still has to check (keyboard, screen reader, content).
No overlay, nothing injected into your site.
3 free scans/day, no signup to see results.
Agencies can put their own name on a white-label PDF for clients.

On the W3C's deliberately broken demo site it found 353 issues across 10 pages.

Founding agencies

I'm looking for 5 founding agencies to shape scheduled re-checks and read-only client links — 50% off for as long as they stay subscribed.

Would love feedback on false positives or confusing fix text: https://a11ykeeper.com

Independent developer. Not a compliance certificate — a triage list.