<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Codex SMS</title>
    <description>The latest articles on DEV Community by Codex SMS (@codexsms).</description>
    <link>https://dev.to/codexsms</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4094385%2Fd461e3e7-6477-49dd-8118-c2753be51efc.png</url>
      <title>DEV Community: Codex SMS</title>
      <link>https://dev.to/codexsms</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/codexsms"/>
    <language>en</language>
    <item>
      <title>Why &amp; How I Built a 100% Browser-Local Bank Statement PDF Reconciler in TypeScript (Zero Backend, $0 Hosting)</title>
      <dc:creator>Codex SMS</dc:creator>
      <pubDate>Sun, 22 Nov 2026 17:15:55 +0000</pubDate>
      <link>https://dev.to/codexsms/why-how-i-built-a-100-browser-local-bank-statement-pdf-reconciler-in-typescript-zero-backend-2efo</link>
      <guid>https://dev.to/codexsms/why-how-i-built-a-100-browser-local-bank-statement-pdf-reconciler-in-typescript-zero-backend-2efo</guid>
      <description>&lt;p&gt;Every tax season, bookkeepers and small business owners hit the same wall: banks lock 12–36 months of historical statements inside PDFs, while existing PDF-to-CSV tools either charge &lt;code&gt;$29/mo+&lt;/code&gt; recurring subscriptions for an annual task or ask you to upload raw financial PII to remote cloud servers.&lt;/p&gt;

&lt;p&gt;Meanwhile, throwing bank statements at an LLM creates a worse problem: &lt;strong&gt;probabilistic math&lt;/strong&gt;. Accounting has an exact ground truth — books either balance to the penny or they don't.&lt;/p&gt;

&lt;p&gt;Here is the architecture behind &lt;strong&gt;Local Bank Statement Reconciler&lt;/strong&gt; (&lt;code&gt;https://vz-001-reconciler.pages.dev&lt;/code&gt;): a pure client-side TypeScript + Vite web app that parses bank PDFs, verifies ledger math deterministically in integer cents, and exports CSV, Excel, and QuickBooks &lt;code&gt;.qbo&lt;/code&gt; (OFX) with &lt;strong&gt;zero backend servers and &lt;code&gt;$0&lt;/code&gt; hosting cost&lt;/strong&gt;.&lt;/p&gt;




&lt;h2&gt;
  
  
  1. Architecture Overview: Why Pure Client-Side Wins on Trust and Cost
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Layer&lt;/th&gt;
&lt;th&gt;Choice&lt;/th&gt;
&lt;th&gt;Why&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;PDF Parsing&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;
&lt;code&gt;pdfjs-dist&lt;/code&gt; (bundled locally)&lt;/td&gt;
&lt;td&gt;Zero external CDN worker requests; raw PDF bytes never leave browser RAM&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Reconciliation&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Deterministic integer-cents engine&lt;/td&gt;
&lt;td&gt;Verifies &lt;code&gt;Opening + Credits - Debits == Closing&lt;/code&gt; with zero IEEE-754 float drift&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Exports&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Pure JS RFC-4180 CSV, OFX &lt;code&gt;.qbo&lt;/code&gt;, SpreadsheetML &lt;code&gt;.xml&lt;/code&gt;
&lt;/td&gt;
&lt;td&gt;Zero server dependencies; works offline&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Licensing&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;WebCrypto ECDSA P-256 (&lt;code&gt;VZ1.&amp;lt;payload&amp;gt;.&amp;lt;sig&amp;gt;&lt;/code&gt;)&lt;/td&gt;
&lt;td&gt;Offline asymmetric signature verification — zero auth server needed&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Hosting &amp;amp; Checkout&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Cloudflare Pages Free + itch.io&lt;/td&gt;
&lt;td&gt;
&lt;code&gt;$0/mo&lt;/code&gt; static hosting with zero-domain external checkout&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Even the built-in &lt;strong&gt;"Try Built-In Sample Statement (&amp;lt;10s)"&lt;/strong&gt; button generates a real PDF 1.4 byte stream &lt;em&gt;in memory&lt;/em&gt; using TypeScript (&lt;code&gt;new Uint8Array(...) -&amp;gt; File&lt;/code&gt;), so a visitor watching &lt;code&gt;F12 -&amp;gt; Network&lt;/code&gt; sees literally &lt;strong&gt;0 HTTP requests&lt;/strong&gt; from click to CSV export.&lt;/p&gt;




&lt;h2&gt;
  
  
  2. Coordinate Clustering &amp;amp; Multi-Line Row Reconstruction
&lt;/h2&gt;

&lt;p&gt;PDFs don't store tables — they store positioned text runs &lt;code&gt;(str, x, y)&lt;/code&gt;. Bank statements complicate this in two ways:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Numeric columns may jitter by &lt;code&gt;1–3 pt&lt;/code&gt; vertically relative to the date/description.&lt;/li&gt;
&lt;li&gt;Long merchant descriptions wrap onto 2–3 continuation lines below the transaction date.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;We normalize every text run through &lt;code&gt;pdfjsLib.Util.transform(viewport.transform, item.transform)&lt;/code&gt;, cluster rows within an adaptive &lt;code&gt;yTol = clamp(lineHeight * 0.45, 2.0, 5.5)&lt;/code&gt;, and merge intra-row glyph spans into column tokens. Once header bands (&lt;code&gt;Date&lt;/code&gt;, &lt;code&gt;Description&lt;/code&gt;, &lt;code&gt;Debit&lt;/code&gt;, &lt;code&gt;Credit&lt;/code&gt;, &lt;code&gt;Amount&lt;/code&gt;, &lt;code&gt;Balance&lt;/code&gt;) establish X-coordinate intervals, any subsequent row that lacks a date and numeric amount is stitched onto the preceding transaction's description rather than dropped or emitted as an orphan row.&lt;/p&gt;




&lt;h2&gt;
  
  
  3. The &lt;code&gt;0&lt;/code&gt;-Glyph Scanned PDF Guard &amp;amp; Integer-Cents Invariant
&lt;/h2&gt;

&lt;p&gt;Two silent failure modes destroy user trust in PDF converters:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Passing an image-only scanned PDF through a text parser -&amp;gt; emitting &lt;code&gt;0&lt;/code&gt; rows -&amp;gt; computing &lt;code&gt;0 + 0 - 0 == 0&lt;/code&gt; -&amp;gt; falsely reporting &lt;code&gt;"Reconciled!"&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Accumulating floating-point dollars (&lt;code&gt;0.1 + 0.2&lt;/code&gt;) and missing penny discrepancies.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Our parser counts selectable non-whitespace glyphs (&lt;code&gt;totalGlyphs&lt;/code&gt;) and inspects &lt;code&gt;page.getOperatorList()&lt;/code&gt; for image XObjects (&lt;code&gt;paintImageXObject&lt;/code&gt;). If &lt;code&gt;totalGlyphs === 0&lt;/code&gt;, the UI halts with an actionable warning explaining why we refuse to upload scanned images to a cloud OCR server. All transaction sums run strictly in integer cents (&lt;code&gt;19.99&lt;/code&gt; -&amp;gt; &lt;code&gt;1999&lt;/code&gt;).&lt;/p&gt;




&lt;h2&gt;
  
  
  4. Offline Asymmetric Licensing via WebCrypto ECDSA P-256
&lt;/h2&gt;

&lt;p&gt;To keep the &lt;code&gt;$19&lt;/code&gt; Pro upgrade 100% offline-capable (no license check-in server, no tracking), the app verifies signed &lt;code&gt;VZ1.&amp;lt;payload_b64u&amp;gt;.&amp;lt;sig_b64u&amp;gt;&lt;/code&gt; tokens locally using &lt;code&gt;window.crypto.subtle&lt;/code&gt; and an embedded ECDSA P-256 &lt;strong&gt;public&lt;/strong&gt; JWK (&lt;code&gt;x&lt;/code&gt; and &lt;code&gt;y&lt;/code&gt; coordinates only — no private scalar &lt;code&gt;d&lt;/code&gt; ever ships in client code).&lt;/p&gt;




&lt;h2&gt;
  
  
  Try It &amp;amp; Inspect the Network Tab
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Live Tool&lt;/strong&gt;: &lt;code&gt;https://vz-001-reconciler.pages.dev&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Test in &amp;lt; 10 Seconds&lt;/strong&gt;: Open &lt;code&gt;F12 -&amp;gt; Network&lt;/code&gt;, click &lt;strong&gt;"Try Built-In Sample Statement"&lt;/strong&gt;, and verify &lt;code&gt;0&lt;/code&gt; network requests from parse to CSV export.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Pricing&lt;/strong&gt;: Core single-statement parsing, full balance check, and up to &lt;code&gt;30&lt;/code&gt; exported CSV rows are &lt;strong&gt;&lt;code&gt;$0&lt;/code&gt; forever (no login)&lt;/strong&gt;. A &lt;strong&gt;&lt;code&gt;$19&lt;/code&gt; one-time Pro license&lt;/strong&gt; unlocks unlimited rows, multi-month PDF merging with duplicate/gap detection, Schedule-C categorization, QuickBooks &lt;code&gt;.qbo&lt;/code&gt; (OFX), Xero CSV, and a multi-sheet Excel audit workbook.&lt;/li&gt;
&lt;/ul&gt;

</description>
      <category>typescript</category>
      <category>webdev</category>
      <category>privacy</category>
      <category>pdf</category>
    </item>
    <item>
      <title>Which Country's Phone Number Works Best for OpenAI/Codex Sign-Up? (2026 Fulfillment Data)</title>
      <dc:creator>Codex SMS</dc:creator>
      <pubDate>Sat, 10 Oct 2026 01:31:02 +0000</pubDate>
      <link>https://dev.to/codexsms/which-countrys-phone-number-works-best-for-openaicodex-sign-up-2026-fulfillment-data-3kod</link>
      <guid>https://dev.to/codexsms/which-countrys-phone-number-works-best-for-openaicodex-sign-up-2026-fulfillment-data-3kod</guid>
      <description>&lt;p&gt;If you're registering for OpenAI's Codex or ChatGPT and need a phone number to complete SMS verification, the common advice on forums is "just pick whatever country is cheapest." The real-world fulfillment data says otherwise: delivery rates for the same platform vary wildly depending on where the number originates.&lt;/p&gt;

&lt;p&gt;This post shares one small operator's records — 31 paid buyers and 100+ activations across 26 countries between August and September 2026 — plus how to interpret them without marketing hype. Sample sizes are small; treat this as directional engineering data, not gospel.&lt;/p&gt;

&lt;h2&gt;
  
  
  Quick answers from fulfillment data
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;US and Japan numbers had the highest paid-buyer delivery rates&lt;/strong&gt; in this dataset: 4 out of 5 buyers received their code after switching to a dedicated US/Japan-focused pool (n=5, small cohort caveat applies). During the earlier broad multi-country period, 15 of 26 buyers completed verification (57.7%).&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Per-activation success&lt;/strong&gt; clustered around the top countries: 

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;United States&lt;/strong&gt;: 26.7% (n=15 attempts, average wait time 164s)&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Japan&lt;/strong&gt;: 25.0% (n=12 attempts, average wait time 487s)&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Colombia&lt;/strong&gt;: 25.0% (n=16 attempts)&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;South Africa&lt;/strong&gt;: 18.2% (n=11 attempts)&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Chile&lt;/strong&gt;: 16.7% (n=12 attempts)&lt;/li&gt;
&lt;/ul&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Portugal performed worst&lt;/strong&gt; in this dataset: only 5.4% across 37 activations. Twenty other countries hovered between 0% and 7%.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Cheap numbers correlate with recycled segments&lt;/strong&gt;: Numbers sold for $0.01–$0.05 on wholesale aggregators failed at the highest rates.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Why "cheapest" is the wrong filter
&lt;/h2&gt;

&lt;p&gt;Public SMS marketplaces sort by price by default, so the top of every list is dominated by numbers from segments that have been queried hundreds of times. OpenAI's fraud and risk models flag those carrier ranges aggressively. &lt;/p&gt;

&lt;p&gt;The result is exactly what the per-activation records show: a long tail of near-zero delivery rates in bargain tiers, and a narrow set of countries where fresh telecommunications prefixes remain unflagged.&lt;/p&gt;

&lt;h3&gt;
  
  
  Practical takeaways from the data:
&lt;/h3&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Prioritize clean prefixes over bargain pricing&lt;/strong&gt;: US and Japan routes consistently showed cleaner reputation scores, even though wholesale unit cost is higher ($0.30–$0.66).&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Stop after two consecutive failures&lt;/strong&gt;: If a country route fails twice, switch regions immediately. Spamming attempts against an already-throttled prefix triggers temporary rate limits on your registration IP.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Account for delayed delivery&lt;/strong&gt;: Japan's average delivery latency was 487 seconds (~8 minutes). Users who cancel after 60 seconds often discard numbers that would have successfully received the code.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Bind authenticator app (TOTP) and recovery email immediately&lt;/strong&gt;: A temporary number is strictly for initial SMS verification, not persistent account recovery. Without immediate TOTP binding, any future session challenge could lock the account.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  The honest caveats
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Sample size&lt;/strong&gt;: 100+ total activations, and the "switched to US/Japan pool" subset rests on 5 buyers. This is empirical operator fulfillment telemetry, not an academic controlled study.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Risk rule volatility&lt;/strong&gt;: OpenAI updates its risk heuristics frequently. What delivered reliably last month can degrade as wholesale segments burn out.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Code delivery != long-term account health&lt;/strong&gt;: Completing phone verification does not exempt an account from IP reputation scoring, fingerprint consistency, or payment gateway checks.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Further Reading &amp;amp; Telemetry
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Full 26-country historical activation breakdown: compiled alongside the co-maintained NumHarbor (numharbor.com) historical reference dataset (as of 2026-10).&lt;/li&gt;
&lt;li&gt;For Chinese-language troubleshooting and step-by-step registration failure fixes, refer to the &lt;a href="https://guide.shanma.store/" rel="noopener noreferrer"&gt;Codex SMS Guide&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;




&lt;p&gt;&lt;em&gt;Disclosure: The author operates Codex SMS (&lt;a href="https://shanma.store" rel="noopener noreferrer"&gt;shanma.store&lt;/a&gt;), where these operational fulfillment logs were recorded. All data is reported as observed without promotional exaggeration.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>openai</category>
      <category>webdev</category>
      <category>security</category>
    </item>
    <item>
      <title>Gmail Sign-Up SMS Issues in 2026: "This Phone Number Cannot Be Used" vs. the Code That Never Arrives</title>
      <dc:creator>Codex SMS</dc:creator>
      <pubDate>Wed, 23 Sep 2026 07:12:35 +0000</pubDate>
      <link>https://dev.to/codexsms/gmail-sign-up-sms-issues-in-2026-this-phone-number-cannot-be-used-vs-the-code-that-never-arrives-2c8f</link>
      <guid>https://dev.to/codexsms/gmail-sign-up-sms-issues-in-2026-this-phone-number-cannot-be-used-vs-the-code-that-never-arrives-2c8f</guid>
      <description>&lt;p&gt;People who get stuck at the phone verification step of Gmail sign-up usually treat it as one problem. It isn't. There are two completely different failure modes, and fixing the wrong one wastes hours:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Google rejects your number before sending anything&lt;/strong&gt; — you get "This phone number cannot be used for verification" or "This phone number has been used too many times for verification."&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Google sends the code, but it never arrives&lt;/strong&gt; — no error message, just silence.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The first is about your environment and number history. The second is about the SMS delivery path. Let's walk both, in troubleshooting order.&lt;/p&gt;

&lt;h2&gt;
  
  
  Quick answers
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Error message about the number itself = &lt;strong&gt;rejection&lt;/strong&gt;: switch to a residential IP, try registering inside the Gmail mobile app, and stop after 2–3 failed attempts (wait 24 hours).&lt;/li&gt;
&lt;li&gt;No error, code just never shows up = &lt;strong&gt;delivery problem&lt;/strong&gt;: check your phone's spam/blocked folders first, then wait 5–10 minutes for international SMS delays.&lt;/li&gt;
&lt;li&gt;If a +86 (mainland China) number keeps getting rejected or codes never arrive, a reliable, low-cost fix is receiving the code once on a &lt;strong&gt;clean overseas number&lt;/strong&gt; that doesn't pass through domestic carrier gateways.&lt;/li&gt;
&lt;li&gt;On the day you register: bind a recovery email, a recovery phone, and app-based 2FA immediately. A temporary number expires — your account shouldn't die with it.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Step 0: Tell the two problems apart
&lt;/h2&gt;

&lt;p&gt;Same wall, different disease. The page behavior tells you which one you have:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;What you see&lt;/th&gt;
&lt;th&gt;What it means&lt;/th&gt;
&lt;th&gt;First move&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;"This phone number cannot be used for verification" / "used too many times"&lt;/td&gt;
&lt;td&gt;Google rejected the number: segment reuse flags or a flagged IP environment&lt;/td&gt;
&lt;td&gt;Switch to residential IP + try the Gmail app; stop after 2–3 failures, wait 24h&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;"Code sent" but nothing arrives&lt;/td&gt;
&lt;td&gt;Delivery problem: carrier blocking, security app filtering, or international gateway delay&lt;/td&gt;
&lt;td&gt;Check spam/blocked folders, wait 5–10 min&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;+86 rejected repeatedly, or codes never arrive long-term&lt;/td&gt;
&lt;td&gt;Number is on a reuse list, or the local delivery path is chronically unstable&lt;/td&gt;
&lt;td&gt;Use a clean overseas number once&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Problem A: Google rejects your number
&lt;/h2&gt;

&lt;p&gt;First, a prerequisite: in some scenarios Google lets you &lt;strong&gt;skip&lt;/strong&gt; phone verification entirely (it depends on your device, registration region, and Google's current risk posture). If you see a "Skip" option, take it and stop reading.&lt;/p&gt;

&lt;p&gt;If there's no skip option, work through this order:&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;1. Clean up your network environment.&lt;/strong&gt; Google is extremely sensitive about registration IPs. Datacenter IPs, heavily shared proxy nodes, and IPs that hop countries within minutes all trigger rejection. Switch to a residential IP, pick one country, stay there for the whole session, then clear browser cache (or switch browser/device) and try once more.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;2. Switch the registration entry point.&lt;/strong&gt; Community testing consistently reports that registering inside the &lt;strong&gt;Gmail mobile app&lt;/strong&gt; succeeds more often than the PC web flow — the app carries device fingerprints and SIM info that make Google's risk engine more lenient. If the web flow keeps rejecting you, try the app path.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;3. Stop hammering the retry button.&lt;/strong&gt; Community reports describe escalating cooldowns: repeated requests from the same number or IP within a short window move you from a few minutes of cooldown, to hours, to a full 24-hour block. Two or three failures = stop, wait 24 hours, change the environment, then retry. Every extra click just extends the cooldown.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;4. Accept that some numbers are burned.&lt;/strong&gt; If the message says "already used for verification too many times," that number is on Google's reuse list. No environment change will save it — you need a different number.&lt;/p&gt;

&lt;h2&gt;
  
  
  Problem B: The code never arrives
&lt;/h2&gt;

&lt;p&gt;Before blaming Google, check your own phone. On Chinese Android phones, the built-in Security Center, harassment-call filters, and spam SMS folders swallow a lot of legitimate verification codes — many people's codes arrived long ago and were auto-filed. On iPhone, check the "Unknown Senders" group in Messages (it exists if you have Filter Unknown Senders enabled in settings).&lt;/p&gt;

&lt;p&gt;If nothing is there, consider international SMS gateway delay: 5–10 minutes is normal, longer at peak hours. Don't request a resend every 30 seconds — the same cooldown logic as above applies.&lt;/p&gt;

&lt;h2&gt;
  
  
  The +86 situation, and the overseas-number fix
&lt;/h2&gt;

&lt;p&gt;For mainland China numbers, both problems stack: +86 segments face higher rejection rates at registration (reuse history), and domestic carriers aggressively filter inbound international SMS. You can fight both fronts separately — or sidestep them by receiving the code once on an overseas number from a clean segment.&lt;/p&gt;

&lt;p&gt;One important distinction: "clean" means a &lt;strong&gt;screened carrier segment without recent abuse flags&lt;/strong&gt;, not "any foreign number." Free public SMS-receiving sites mostly hand out VoIP segments that Google's filters recognize instantly, and shared numbers on those sites are visible to everyone — anyone can use the same number to recover &lt;em&gt;your&lt;/em&gt; account later. That's the real risk of free platforms, worse than the low success rate.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Disclosure: the team behind this article operates a pay-per-use SMS verification service, Codex SMS (&lt;a href="https://shanma.store/" rel="noopener noreferrer"&gt;shanma.store&lt;/a&gt;), that assigns US (+1) numbers for Google registration, with Japan (+81) as fallback. Same-team interest — take this paragraph as an ad if you like; the troubleshooting steps above stand on their own. Check the site for current pricing.&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The day you register: three things to do immediately
&lt;/h2&gt;

&lt;p&gt;A temporary number expires after it does its job. But Gmail may ask for re-verification later — new device, unusual login, security check. Registration success is only halftime. Same day, do these three:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Bind a recovery email&lt;/strong&gt; — one you control long-term. This is the most important safety net once the temporary number dies.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Set a recovery phone number&lt;/strong&gt; — your own +86 number generally works here (community experience: reuse limits target registration, not recovery). If your +86 chronically misses international SMS, treat the recovery email as the primary fallback and consider the voice-call option. This demotes "phone number" from your only line of defense to one of several.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Turn on 2-Step Verification with an authenticator app (TOTP)&lt;/strong&gt; — not SMS. TOTP doesn't depend on any phone number, so a dead temporary number can't lock you out.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Worth remembering: a Gmail account is also your passport to YouTube, Google Play, and Drive — and it's a common sign-in option for AI services like OpenAI and Claude. Treat it as infrastructure, not a throwaway.&lt;/p&gt;

&lt;h2&gt;
  
  
  FAQ
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Is my +86 number banned if Gmail codes never arrive?&lt;/strong&gt;&lt;br&gt;
Not banned. It's usually one of the two problems above: Google rejected the number before sending (IP environment or segment history), or the code was sent and got filtered by your carrier's anti-fraud SMS blocking, a security app, or gateway delay. Check filtered folders before assuming anything.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What if it says "this phone number has been used too many times for verification"?&lt;/strong&gt;&lt;br&gt;
That number (or its segment) has already registered Google accounts and hit the reuse limit. Retrying is pointless. Use a number that has never registered a Google account — your own backup SIM, a family member's, or a clean overseas segment.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Does registering through the Gmail app really work better than the web?&lt;/strong&gt;&lt;br&gt;
Community testing says yes, consistently. The app carries device fingerprints and SIM info, so risk scoring is more lenient. It works best combined with a fixed-country residential IP.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is a Gmail account registered with a temporary number safe?&lt;/strong&gt;&lt;br&gt;
The number is only used to receive one code at registration. Account safety depends on what you do afterward: recovery email, recovery phone, and TOTP on day one. What you should avoid are shared numbers on free public platforms — the number is visible to everyone, and anyone can use it to recover your account.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Do I still need to bind my own phone number afterward?&lt;/strong&gt;&lt;br&gt;
Recommended. Temporary numbers expire. Bind a long-term recovery phone if you have a reliable one (recovery email stays the primary fallback), plus TOTP — multiple layers, and the account is truly yours.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Based on public documentation and community cases as of September 2026. Google's risk rules change frequently; the live prompts on the registration page are the final word. The authors have no affiliation with Google.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>gmail</category>
      <category>tutorial</category>
      <category>verification</category>
      <category>google</category>
    </item>
    <item>
      <title>Discord Phone Verification in 2026: Why +86 and VoIP Numbers Get Rejected (and What Actually Works)</title>
      <dc:creator>Codex SMS</dc:creator>
      <pubDate>Sun, 20 Sep 2026 08:01:33 +0000</pubDate>
      <link>https://dev.to/codexsms/discord-phone-verification-in-2026-why-86-and-voip-numbers-get-rejected-and-what-actually-works-159h</link>
      <guid>https://dev.to/codexsms/discord-phone-verification-in-2026-why-86-and-voip-numbers-get-rejected-and-what-actually-works-159h</guid>
      <description>&lt;p&gt;If you've hit Discord's phone verification wall, you already know the two error messages by heart: the one where your country's code isn't in the list at all, and the one where the number you pasted gets rejected as invalid. Both are real restrictions, not bugs — and most tutorials we found conflate the two.&lt;/p&gt;

&lt;p&gt;This is what we've observed from community reports and our own testing in 2026. One disclosure up front: our team runs an SMS-verification service, which we'll mention where relevant — we've flagged it clearly rather than hiding it.&lt;/p&gt;

&lt;h2&gt;
  
  
  When Discord asks for a phone number
&lt;/h2&gt;

&lt;p&gt;Discord triggers phone verification in a lot of situations:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Sign-up&lt;/strong&gt; — in some regions and network environments, verification is required before you can finish registering.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;New device or new IP&lt;/strong&gt; — same pattern as every major platform: environment change triggers a second check.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Sensitive actions&lt;/strong&gt; — changing account details, joining certain servers, or using specific features.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Risk-control flags&lt;/strong&gt; — behavior that looks automated (rapid IP switching, bulk patterns) forces verification.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The important part: these triggers &lt;strong&gt;recur&lt;/strong&gt;. Passing verification once does not mean you're done.&lt;/p&gt;

&lt;h2&gt;
  
  
  The two walls
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Wall 1: +86 numbers aren't supported.&lt;/strong&gt; Discord's verification simply doesn't include mainland China's +86 range. This is a product-level regional restriction. No tutorial about "receiving Discord SMS on +86" can fix this layer — the correct move is to switch to another country's number range entirely.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Wall 2: VoIP/virtual numbers get rejected.&lt;/strong&gt; Even with an overseas number, Discord blocks many VoIP ranges. This is why free public SMS-receiving sites almost never work for Discord — their numbers are nearly all VoIP. The diagnostic is simple: retry with a number from a real carrier's range. If it goes through, it was a number-range problem. If it doesn't, look at your account status instead.&lt;/p&gt;

&lt;p&gt;Stack those two walls together and the whole "Discord verification is hard" story reduces to one sentence: &lt;strong&gt;you need a non-+86 number on a real carrier's range.&lt;/strong&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Three paths, compared
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Path&lt;/th&gt;
&lt;th&gt;Barrier&lt;/th&gt;
&lt;th&gt;Cost&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Your own overseas SIM&lt;/td&gt;
&lt;td&gt;Need the physical card&lt;/td&gt;
&lt;td&gt;Monthly fee&lt;/td&gt;
&lt;td&gt;Most stable; the number stays yours long-term&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Borrow a friend's overseas number&lt;/td&gt;
&lt;td&gt;Need the connection&lt;/td&gt;
&lt;td&gt;Free&lt;/td&gt;
&lt;td&gt;Careful: the number binds to the account and is needed again for recovery&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Prepaid verification services&lt;/td&gt;
&lt;td&gt;Sign-up only&lt;/td&gt;
&lt;td&gt;≈US$0.50–1 per use&lt;/td&gt;
&lt;td&gt;You redeem a code and get a temporary real-carrier number; look for ones that replace the number free if no code arrives&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The selection logic: &lt;strong&gt;long-term use → real SIM; one-time need → a prepaid service.&lt;/strong&gt; (The third row is the category our own product belongs to, so weigh it accordingly.)&lt;/p&gt;

&lt;h2&gt;
  
  
  Why some countries' ranges work better than others
&lt;/h2&gt;

&lt;p&gt;SMS-verification providers track success volume by service × country. In the operational data we see (September 2026), Brazilian ranges currently lead for Discord, with Indonesia, Morocco, Thailand, the Philippines, and South Africa as common backups. This isn't superstition — high recent success volume means large numbers of users are passing Discord verification with those ranges &lt;em&gt;right now&lt;/em&gt;. It's the best available signal of current deliverability, though not a permanent guarantee.&lt;/p&gt;

&lt;p&gt;When picking a number, "real carrier range + high recent success volume" beats blind retrying every time.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;(As disclosed above: our team's Discord verification service at &lt;a href="https://shanma.store/" rel="noopener noreferrer"&gt;shanma.store&lt;/a&gt; allocates ranges by exactly this principle — Brazil first, five-country backup pool.)&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  After you pass: reducing recurrence
&lt;/h2&gt;

&lt;p&gt;One honest boundary first: &lt;strong&gt;a temporary number solves this verification only&lt;/strong&gt; — it can't receive codes for you in the future. If you plan to keep the account, the most robust structure is:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Enable TOTP two-factor (authenticator app) right after verifying.&lt;/strong&gt; It's Discord's own official mechanism and reduces how often login/device checks lean on your phone. It won't exempt you from server-level phone gates (some servers require a verified phone regardless of 2FA) or from future re-checks — but it matters more than any number-picking strategy.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Keep your login environment stable&lt;/strong&gt; — same device, same IP range. Environment changes are the top trigger.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Save your verification record&lt;/strong&gt; — a screenshot of the number and the time. You'll want it for re-binding, recovery, or support appeals.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Don't hammer the resend button.&lt;/strong&gt; If the code hasn't arrived, wait 2–3 minutes first. If two rounds produce nothing, then switch numbers — rapid retries are wasteful and can trip risk flags.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  FAQ
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Can Discord be verified with a +86 number?&lt;/strong&gt;&lt;br&gt;
No. The +86 range isn't in Discord's verification list. You need another country's range.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Discord rejected my virtual number. Now what?&lt;/strong&gt;&lt;br&gt;
Retry with a real carrier's range. Free public SMS sites are mostly VoIP and rejection is expected.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Will Discord ask again after I verify once?&lt;/strong&gt;&lt;br&gt;
Yes — new devices, new IPs, and sensitive actions can all re-trigger it. Stable environment + TOTP is the durable answer.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The number I got was "already in use" — why?&lt;/strong&gt;&lt;br&gt;
Temporary numbers get recycled. If a number was previously bound to another Discord account, Discord rejects it — request a replacement number rather than retrying the same one.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is there account risk in verifying with a temporary number?&lt;/strong&gt;&lt;br&gt;
Yes, be honest with yourself about this. Discord actively blocks VoIP ranges, allows only one account per number, and does enforce against shared/temporary numbers — accounts verified this way can be limited later. Verifying and then immediately enabling TOTP plus normal usage behavior is the best mitigation available.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Written September 2026 from community reports and hands-on observation. Discord's risk policies and supported ranges change without notice — trust the live prompts on Discord's own pages over any static guide, including this one.&lt;/em&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Automation note (2026-11-16): If you test verification flows from Python 3.10+ CI harnesses, our open-source zero-dependency &lt;a href="https://github.com/thirddiver/shanma-sms-sdk" rel="noopener noreferrer"&gt;&lt;code&gt;codex-sms&lt;/code&gt; SDK (&lt;code&gt;v0.3.3&lt;/code&gt;)&lt;/a&gt; handles single-use guest CDK redemption (&lt;code&gt;from codex_sms import CodexSmsClient&lt;/code&gt;), monotonic OTP polling, and HTTP 409 abandon-SMS guards.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>discord</category>
      <category>tutorial</category>
    </item>
    <item>
      <title>OpenAI Phone Verification in 2026: ChatGPT vs Codex vs API Keys (What Actually Needs a Number)</title>
      <dc:creator>Codex SMS</dc:creator>
      <pubDate>Tue, 25 Aug 2026 15:09:57 +0000</pubDate>
      <link>https://dev.to/codexsms/openai-phone-verification-in-2026-chatgpt-vs-codex-vs-api-keys-what-actually-needs-a-number-1c97</link>
      <guid>https://dev.to/codexsms/openai-phone-verification-in-2026-chatgpt-vs-codex-vs-api-keys-what-actually-needs-a-number-1c97</guid>
      <description>&lt;p&gt;If you search "OpenAI SMS verification" you will find a lot of outdated tutorials. The rules changed, and they are not the same across OpenAI's products. Here is the current state as of August 2026, based on OpenAI's official help docs.&lt;/p&gt;

&lt;h2&gt;
  
  
  The three scenarios are different
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;1. Creating a new ChatGPT account — no phone number required anymore.&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;OpenAI's help documentation now states that creating an OpenAI account and using ChatGPT no longer requires phone verification. If a tutorial tells you to get a phone number &lt;em&gt;before&lt;/em&gt; signing up for ChatGPT, it is outdated.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;2. Logging in or using Codex — phone verification IS triggered.&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;When you sign in to Codex (the CLI, the IDE extension, or the web version), you will most likely be asked to complete a one-time phone verification — even if your account never verified a phone number when registering ChatGPT. This is tied to Codex's stricter risk-control policy, and it catches many users off guard.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;3. Generating your first API key — still required.&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;On platform.openai.com, phone verification remains a mandatory step when you create your &lt;strong&gt;first&lt;/strong&gt; API key. Subsequent keys do not require re-verification.&lt;/p&gt;

&lt;p&gt;So before following any guide, figure out which of the three pages you are actually looking at. The fix for each is different.&lt;/p&gt;

&lt;h2&gt;
  
  
  What the verification page looks like
&lt;/h2&gt;

&lt;p&gt;Depending on your country/region, the official verification screen may offer to send the one-time code via &lt;strong&gt;SMS or WhatsApp&lt;/strong&gt;. A few ground rules that apply no matter what:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Only use a number you are authorized to use, can control long-term, and that complies with OpenAI's rules.&lt;/li&gt;
&lt;li&gt;Never share the one-time code with anyone.&lt;/li&gt;
&lt;li&gt;A one-time code is not an account recovery method.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  When something goes wrong
&lt;/h2&gt;

&lt;p&gt;Number rejected, code never arrives, region not supported? Keep the on-screen error message, cross-check OpenAI's help center and the list of supported regions, and go through official support if needed. Repeatedly trying numbers of unknown origin, buying pre-made accounts, or changing your region/payment profile raises your account and payment risk significantly.&lt;/p&gt;

&lt;h2&gt;
  
  
  Don't build important accounts on one-time codes
&lt;/h2&gt;

&lt;p&gt;Whether or not your verification succeeds, any account that matters should have a recovery email you control long-term, an authenticator app, or another officially recognized backup method. Temporary numbers expire. Other people's numbers disappear. Neither can replace proper account hygiene.&lt;/p&gt;

&lt;h2&gt;
  
  
  Handling Automated QA &amp;amp; Stateful Verification Safely
&lt;/h2&gt;

&lt;p&gt;In developer test suites where an isolated one-time verification corridor is needed (such as CLI / IDE onboarding tests or first API key generation), treat verification as a bounded state machine rather than a blind retry loop (&lt;code&gt;time.monotonic()&lt;/code&gt; deadlines, &lt;code&gt;&amp;gt;= 4s&lt;/code&gt; poll intervals, and explicit &lt;code&gt;HTTP 409 Conflict&lt;/code&gt; checks before swapping unresponsive numbers). Full disclosure: we maintain &lt;a href="https://shanma.store/" rel="noopener noreferrer"&gt;Codex SMS&lt;/a&gt; and its zero-dependency MIT Python SDK (&lt;a href="https://github.com/thirddiver/shanma-sms-sdk" rel="noopener noreferrer"&gt;&lt;code&gt;codex-sms&lt;/code&gt;&lt;/a&gt;):&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="k"&gt;import&lt;/span&gt; &lt;span class="n"&gt;os&lt;/span&gt;
&lt;span class="k"&gt;from&lt;/span&gt; &lt;span class="n"&gt;codex_sms&lt;/span&gt; &lt;span class="k"&gt;import&lt;/span&gt; &lt;span class="nc"&gt;CodexSmsClient&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nc"&gt;ActivationNotFoundError&lt;/span&gt;

&lt;span class="n"&gt;client&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="nc"&gt;CodexSmsClient&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;cdkey&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="n"&gt;os&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;environ&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"CODEX_SMS_CDK"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;timeout&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;30&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="k"&gt;with&lt;/span&gt; &lt;span class="n"&gt;client&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;activation&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="k"&gt;as&lt;/span&gt; &lt;span class="n"&gt;act&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
    &lt;span class="n"&gt;sms&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;act&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;wait_for_sms&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;timeout&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;120&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;interval&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;5&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
    &lt;span class="k"&gt;if&lt;/span&gt; &lt;span class="ow"&gt;not&lt;/span&gt; &lt;span class="n"&gt;sms&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
        &lt;span class="n"&gt;act&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;switch&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
        &lt;span class="n"&gt;sms&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;act&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;wait_for_sms&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;timeout&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;120&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;interval&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;5&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;To be honest about the limits: a one-time number works only for that single verification, success depends on number-pool stock and the target platform's real-time risk control, and nobody can guarantee 100% delivery. It is also not a long-term recovery method for your account.&lt;/p&gt;

&lt;h2&gt;
  
  
  References
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;a href="https://help.openai.com/" rel="noopener noreferrer"&gt;OpenAI help center — phone verification&lt;/a&gt; (check the current page; rules change)&lt;/li&gt;
&lt;li&gt;My detailed Chinese-language notes on the ChatGPT/Codex/API differences: &lt;a href="https://guide.shanma.store/news/openai-sms-verification.html" rel="noopener noreferrer"&gt;guide.shanma.store&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;em&gt;Disclosure: I am affiliated with shanma.store. This post describes official rules and does not provide methods to bypass or evade verification.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>openai</category>
      <category>webdev</category>
    </item>
  </channel>
</rss>
