<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[Building a voice-follow teleprompter on iPhone without sending audio anywhere]]></title><description><![CDATA[Building a voice-follow teleprompter on iPhone without sending audio anywhere]]></description><link>https://agnas.hashnode.dev</link><generator>RSS for Node</generator><lastBuildDate>Wed, 18 Nov 2026 14:53:27 GMT</lastBuildDate><atom:link href="https://agnas.hashnode.dev/rss.xml" rel="self" type="application/rss+xml"/><language><![CDATA[en]]></language><ttl>60</ttl><item><title><![CDATA[I audited the top 20 iPhone teleprompter apps: they have 30,000+ ratings, I have 6]]></title><description><![CDATA[What public data shows about the top 20 iPhone teleprompter apps: ratings, where their traffic comes from, and what a small app can do.]]></description><link>https://agnas.hashnode.dev/i-audited-the-top-20-iphone-teleprompter-apps-they-have-30-000-ratings-i-have-6</link><guid isPermaLink="true">https://agnas.hashnode.dev/i-audited-the-top-20-iphone-teleprompter-apps-they-have-30-000-ratings-i-have-6</guid><category><![CDATA[iOS]]></category><category><![CDATA[marketing]]></category><category><![CDATA[SEO]]></category><category><![CDATA[startup]]></category><dc:creator><![CDATA[Andrew Gnasevych]]></dc:creator><pubDate>Wed, 18 Nov 2026 12:20:54 GMT</pubDate><content:encoded><![CDATA[<p>I make a small iPhone teleprompter app and it is nowhere in search for the word "teleprompter". Before writing more pages, I wanted to see who is there and where their traffic comes from. This is what public data shows, with the gaps marked.</p>
<p>Disclosure: the app is mine, <a href="https://apps.apple.com/ua/app/teleprompter-offline-private/id6808964093?pt=128162591&amp;ct=hn_competitors&amp;mt=8">Teleprompter Offline &amp; Private</a>. I only used public pages and the iTunes Search API on 15 November. Traffic numbers for other companies are third-party estimates and the tools disagree with each other.</p>
<h2>Where I stand</h2>
<p>I pulled the top 20 for 11 searches (US, Brazil, Mexico, Germany). My app appears in exactly one: #1 for "teleprompter offline" in the US. For plain "teleprompter" it is absent in all four countries, and absent for "ai teleprompter", "voice scroll teleprompter" and "teleprompter for video" in the US.</p>
<h2>Who is on top (US, "teleprompter")</h2>
<ul>
<li>Teleprompter Apps Ltd: five apps in the top 13 (#1, #6, #7, #8, #13). Their top app has about 34,900 ratings.</li>
<li>Teleprompter for Video (Norton Five): #2, about 15,400 ratings.</li>
<li>Teleprompter.com: #3, about 30,400 ratings.</li>
<li>BIGVU: #4, about 36,500 ratings.</li>
<li>PromptSmart Pro: #10, about 12,000 ratings, a one-time $29.99.</li>
</ul>
<p>I have 6 ratings. Ratings count is the gap I cannot close with a landing page.</p>
<h2>Where their traffic seems to come from</h2>
<ol>
<li>App Store search, for nearly all of them. Several have a website that is only a storefront (Teleprompter for Video has 7 URLs in its sitemap).</li>
<li>SEO content, but only one does it seriously: Teleprompter.com has about 720 URLs in its sitemap, a blog, a web version and an llms.txt file for AI search. The company says 200,000+ monthly users and 5M+ downloads (its own claim); AppGoblin estimates 3.8M installs.</li>
<li>Social, but only one is big: BIGVU has about 29,000 followers on TikTok (684 videos) and 23,000 on YouTube, runs paid TikTok ads, and a third-party listing says it pays affiliates 20% recurring. Third-party tools put its site at roughly 85,000 visits a month, 60 to 90 percent of it direct. Teleprompter.com has 3,300 YouTube subscribers and 546 on TikTok. PromptSmart has 850 on YouTube.</li>
<li>App-store coverage. Teleprompter Apps Ltd just fills the results with several apps for the same query.</li>
</ol>
<h2>What I take from it</h2>
<ul>
<li>I will not fight Teleprompter.com for "teleprompter online" or "free teleprompter". I write narrow pages instead: offline, voice-follow, mirror mode, "for video", and local-language versions.</li>
<li>Ratings are the real barrier, so I shipped a soft rating prompt (it asks "Do you like it?" first and only then opens the system dialog).</li>
<li>Social is mostly empty in this niche outside BIGVU, so short videos should stand out more than they would in a crowded niche.</li>
<li>Every page and post gets its own campaign token, so I can see which one sends App Store visits and not only impressions.</li>
</ul>
<h2>What I do not know</h2>
<p>Exact traffic for any of these companies. The iTunes API order is not the same as the live App Store ranking. I did not check Instagram or Facebook numbers.</p>
<p>If you run a small app in a crowded category, which did more for you: a niche keyword or the first hundred ratings?</p>
<p>Links: <a href="https://www.a-devgroup.com/teleprompter/best-teleprompter-apps/">six teleprompter apps compared</a>, <a href="https://www.a-devgroup.com/teleprompter/">teleprompter pages</a> and <a href="https://agnas.hashnode.dev/keyword-research-for-an-iphone-app-in-20-countries-using-only-google-trends-and-search-console">how I picked keywords with Google Trends</a>.</p>
]]></content:encoded></item><item><title><![CDATA[Keyword research for an iPhone app in 20 countries, using only Google Trends and Search Console]]></title><description><![CDATA[What Google Trends, Search Console and App Store Connect showed for a teleprompter app in 20 locales, with the real numbers and what I changed.]]></description><link>https://agnas.hashnode.dev/keyword-research-for-an-iphone-app-in-20-countries-using-only-google-trends-and-search-console</link><guid isPermaLink="true">https://agnas.hashnode.dev/keyword-research-for-an-iphone-app-in-20-countries-using-only-google-trends-and-search-console</guid><category><![CDATA[SEO]]></category><category><![CDATA[iOS]]></category><category><![CDATA[marketing]]></category><category><![CDATA[analytics]]></category><dc:creator><![CDATA[Andrew Gnasevych]]></dc:creator><pubDate>Sun, 15 Nov 2026 14:04:19 GMT</pubDate><content:encoded><![CDATA[<p>I make a small iPhone teleprompter app. It has App Store listings in 25 languages, but my website pages were written on guesses. Last week I checked those guesses against data: Google Trends for 20 countries, Search Console for my site and App Store Connect for the app. Here is what I found, with the numbers.</p>
<p>Disclosure: the app is mine, <a href="https://apps.apple.com/ua/app/teleprompter-offline-private/id6808964093?pt=128162591&amp;ct=hn_trends20&amp;mt=8">Teleprompter Offline &amp; Private</a>. Trends numbers are relative indexes inside one comparison, not search volumes, and they cannot be compared between countries.</p>
<h2>Where the downloads come from</h2>
<p>App Store Connect, 2 October to 14 November, 108 downloads:</p>
<table>
<thead>
<tr>
<th>Source</th>
<th>Downloads</th>
</tr>
</thead>
<tbody><tr>
<td>App Store search</td>
<td>64</td>
</tr>
<tr>
<td>Other apps (referrer)</td>
<td>36</td>
</tr>
<tr>
<td>App Store browse</td>
<td>6</td>
</tr>
<tr>
<td>Web</td>
<td>2</td>
</tr>
</tbody></table>
<p>The web brought 2 downloads out of 108, so for this app the App Store listing matters far more than the website. By territory the top ten were US 12, Saudi Arabia 11, Ukraine 11, Philippines 10, Brazil 8, China 5, UK 5, Spain 4, Germany 4, Nigeria 3. Small numbers, but they told me which locales to look at.</p>
<h2>What Trends showed</h2>
<p>I compared five phrases per country, in the local language, over 12 months.</p>
<p><strong>1. The plain word wins.</strong> "teleprompter" scored 84 in Brazil, 63 in Indonesia, 61 in Spain, 52 in Mexico, 48 in Germany and 45 in Italy, while "teleprompter gratis", "grátis" and "kostenlos" scored 0 to 3. Three of my pages targeted those local "free" words. Trends hides very low volumes, so this does not prove zero searches, but it shows I was aiming at the thin end.</p>
<p><strong>2. The long tail is English, even abroad.</strong> The top related queries in most countries were English: "teleprompter online", "teleprompter online free", "teleprompter for video", "free teleprompter app", "teleprompter app iphone". In Brazil the local word that showed up was "gratuito", not "grátis".</p>
<p><strong>3. My US pages missed.</strong> In the US, "teleprompter app" scored 61, "free teleprompter" 40 and "teleprompter online" 23. "voice activated teleprompter" scored 0 and "teleprompter for youtube" 1, and those were two of my pages.</p>
<p><strong>4. Some countries use another word.</strong> In France "prompteur" scored 62 against 21 for "teleprompter". In Turkey "prompter" scored 56 against 21, and in Poland 32 against 14.</p>
<p><strong>5. A competitor is breaking out.</strong> "cueprompter", a free browser teleprompter, was a breakout related query in 7 of the 20 countries (US, Mexico, Spain, France, Turkey, Indonesia, Brazil).</p>
<p>Arabic, Hebrew, Thai, Korean and Chinese queries in their own scripts were below Trends' threshold. People there search in English or not at all, so those languages need a good App Store listing, not web pages.</p>
<h2>What Search Console said</h2>
<p>For a-devgroup.com, 18 October to 15 November: 446 impressions, 4 clicks, average position 12.3. 148 of the impressions went to the home page and 86 to the brand query "devgroup".</p>
<p>The best non-brand result was not an app page. It was a small speaking-time calculator: 47 impressions at position 8.1, from queries like "how long does it take to speak 150 words". The new teleprompter and comparison pages I added from 9 November mostly have 0 to 2 impressions so far, and one old comparison page has 8. They are days old, so I read this as "not measured yet", not as a result.</p>
<p>One thing I want to be straight about: impressions rose from about 8 a day to about 30 a day on 4 November, five days before the new pages existed. I cannot credit the pages for that jump.</p>
<h2>What I changed</h2>
<ul>
<li>Renamed my "for YouTube" page to "teleprompter for video", because "for video" was rising in several countries and "for youtube" was near zero.</li>
<li>Rewrote the Brazilian page around "gratuito" and added the "online" angle.</li>
<li>Added pages for "free teleprompter app", "CuePrompter alternative" and "online teleprompter vs app", using only facts from CuePrompter's own site and saying that its maker also sells apps.</li>
<li>Built a new page for "how many words in a 5 minute speech", the one cluster where Search Console already showed demand.</li>
</ul>
<p>I will check impressions per page in three to four weeks and write down what moved. If you did similar research for a tiny product, which source of keyword data did you trust most?</p>
<p>Links: <a href="https://www.a-devgroup.com/teleprompter/">teleprompter pages</a>, <a href="https://www.a-devgroup.com/how-many-words-in-a-5-minute-speech/">the speaking-time page</a>.</p>
]]></content:encoded></item><item><title><![CDATA[What iPhone teleprompter apps tell Apple about you (privacy labels compared)]]></title><description><![CDATA[I compared the App Store privacy labels of six iPhone teleprompter apps: who tracks you, what is linked to you and what stays on the phone.]]></description><link>https://agnas.hashnode.dev/what-iphone-teleprompter-apps-tell-apple-about-you-privacy-labels-compared</link><guid isPermaLink="true">https://agnas.hashnode.dev/what-iphone-teleprompter-apps-tell-apple-about-you-privacy-labels-compared</guid><category><![CDATA[iOS]]></category><category><![CDATA[privacy]]></category><category><![CDATA[mobile]]></category><category><![CDATA[Apple]]></category><dc:creator><![CDATA[Andrew Gnasevych]]></dc:creator><pubDate>Thu, 12 Nov 2026 10:22:49 GMT</pubDate><content:encoded><![CDATA[<p>A teleprompter app listens to you read, sees your face if it records video, and holds your scripts. That makes it a good test of what an app does with personal data. I compared the App Store privacy labels of six popular iPhone teleprompters. Disclosure: I build one of them, Teleprompter Offline &amp; Private.</p>
<p><em>Data: App Store privacy labels, US storefront, checked on 9 November 2026. Labels change, so check the listing.</em></p>
<p><strong>How to read the labels.</strong> Apple splits them into three groups. "Data Used to Track You" means data used across apps and websites owned by other companies. "Data Linked to You" means data tied to your identity. "Data Not Linked to You" means it is collected but not tied to you. Labels are filled in by the developers themselves, so treat them as claims, not audits.</p>
<p><strong>The comparison</strong></p>
<ul>
<li><strong>Teleprompter Offline &amp; Private</strong> (mine): no tracking. Not linked to you: Identifiers, Usage Data, Diagnostics.</li>
<li><strong>PromptSmart Pro</strong>: no tracking. Everything it lists is not linked to you: Contact Info, User Content, Identifiers, Usage Data, Diagnostics.</li>
<li><strong>Teleprompter (Teleprompter Apps Ltd)</strong>: no tracking. Linked to you: Contact Info, User Content, Identifiers. Not linked: Usage Data.</li>
<li><strong>Teleprompter for Video (Norton Five)</strong>: tracking: Contact Info. Linked to you: Purchases, Contact Info, User Content. Not linked: Diagnostics.</li>
<li><a href="http://Teleprompter.com"><strong>Teleprompter.com</strong></a>: tracking: Identifiers. Linked to you: Contact Info, Identifiers, Usage Data.</li>
<li><strong>BIGVU</strong>: tracking: Identifiers, Usage Data. Linked to you: Contact Info, User Content, Identifiers, Usage Data. Not linked: Diagnostics.</li>
</ul>
<p><strong>What stands out</strong></p>
<ol>
<li>Only two of the six, PromptSmart Pro and mine, report nothing linked to the user and no tracking.</li>
<li>Three of the six report some form of tracking. For a tool you use to read a script, that is a lot.</li>
<li>"User Content" linked to you appears in three apps. For a teleprompter that likely means your scripts or recordings sit in an account. That is a feature if you want cloud sync, and a cost if you do not.</li>
<li>Mine is not empty either. I collect anonymous identifiers, usage and diagnostics, and I report them as not linked to you. "Private" here means no account, no tracking and speech recognition on the phone, not "collects nothing".</li>
</ol>
<p><strong>Why the labels differ</strong></p>
<p>An account is the main driver. Apps with sign-in tend to link contact info and identifiers to you. Apps with cloud scripts link user content. An app that works offline without an account has nothing to link.</p>
<p><strong>How to check any app in 20 seconds</strong></p>
<p>Open its App Store page, scroll to "App Privacy" and read the three groups. If an app needs an account for a basic job like reading a script, the label will say so.</p>
<p>Side-by-side feature comparison of the same apps: <a href="https://www.a-devgroup.com/teleprompter/best-teleprompter-apps/">https://www.a-devgroup.com/teleprompter/best-teleprompter-apps/</a></p>
<p>My app's page: <a href="https://www.a-devgroup.com/teleprompter/">https://www.a-devgroup.com/teleprompter/</a></p>
<p>My app on the App Store: <a href="https://apps.apple.com/ua/app/teleprompter-offline-private/id6808964093?pt=128162591&amp;ct=hn_privlabels&amp;mt=8">https://apps.apple.com/ua/app/teleprompter-offline-private/id6808964093?pt=128162591&amp;ct=hn_privlabels&amp;mt=8</a></p>
]]></content:encoded></item><item><title><![CDATA[Building a voice-follow teleprompter on iPhone without sending audio anywhere]]></title><description><![CDATA[A teleprompter that scrolls as you speak sounds simple: recognize speech, find the words in the script, move the text. I built that for iOS and the…]]></description><link>https://agnas.hashnode.dev/building-a-voice-follow-teleprompter-on-iphone-without-sending-audio-anywhere</link><guid isPermaLink="true">https://agnas.hashnode.dev/building-a-voice-follow-teleprompter-on-iphone-without-sending-audio-anywhere</guid><dc:creator><![CDATA[Andrew Gnasevych]]></dc:creator><pubDate>Mon, 09 Nov 2026 16:31:02 GMT</pubDate><content:encoded><![CDATA[<p>A teleprompter that scrolls as you speak sounds simple: recognize speech, find the words in the script, move the text. I built that for iOS and the recognizing part was the easy part. The hard parts were making the scroll feel calm and not trusting the recognizer too much. Here is what I learned. Disclosure: this is my app, Teleprompter Offline &amp; Private.</p>
<p><strong>1. Keep recognition on the device</strong></p>
<p>I use Apple's speech recognition and require on-device mode. If the language has no on-device model, the app says so instead of silently falling back to a server. That choice makes the app work in Airplane Mode and means your voice never leaves the phone. The cost: some languages need the offline speech pack, downloaded once in iOS settings.</p>
<p><strong>2. Match words, not text</strong></p>
<p>The recognizer returns a stream of partial transcripts. I align them against the script with a fuzzy matcher that searches a lookahead window of 40 to 70 words past the current position. Fuzzy, because people skip words, say "gonna" for "going to", and the recognizer mishears.</p>
<p><strong>3. The bug that looked like a smoothing problem</strong></p>
<p>The scroll jerked. I spent days tuning easing curves. Then I found the cause was the matcher, not the animation. One stray match on a breath or a noise could jump the cursor across the whole lookahead window, 40 words at once, and the text lurched.</p>
<p>The fix was boring: cap how far the cursor can advance per update (12 words). If the speaker really did jump ahead, the next updates catch up. A jump that fast is almost always noise.</p>
<p><strong>4. Start every animation from zero velocity</strong></p>
<p>My first smoothing carried velocity from one animation segment to the next. It looked elegant and sometimes overshot the target wildly after a noisy update. Now every segment starts at zero velocity and eases toward the target with a critically damped curve. From zero velocity a critically damped ease approaches the target monotonically and never crosses it, so the text cannot overshoot.</p>
<p><strong>5. Plan for the recognizer going quiet</strong></p>
<p>During video recording the recognizer can lose the audio stream. If the stream stops for more than two seconds, the app falls back to a steady words-per-minute scroll, keeps recording, and tells the user. A teleprompter that stops dead on camera is worse than one that scrolls at the wrong speed.</p>
<p><strong>6. Rewind is rarer than you think</strong></p>
<p>Going back needs a stricter rule: the cursor must have been standing still for a couple of updates before I allow a relocation, and a forward jump must match at least four words. Otherwise repeated phrases in a script pull the cursor to the wrong place.</p>
<p><strong>What I would tell myself at the start</strong></p>
<ul>
<li>Put the limits in the matcher, not in the animation.</li>
<li>Prefer a rule you can prove (monotonic approach to target) over a clever one.</li>
<li>Test with a script that repeats phrases. Real scripts do.</li>
</ul>
<p>The app is free to download, works offline and needs no account: <a href="https://apps.apple.com/ua/app/teleprompter-offline-private/id6808964093?pt=128162591&amp;ct=hn_voicefollow&amp;mt=8">https://apps.apple.com/ua/app/teleprompter-offline-private/id6808964093?pt=128162591&amp;ct=hn_voicefollow&amp;mt=8</a> Comparison with other teleprompter apps: <a href="https://www.a-devgroup.com/teleprompter/best-teleprompter-apps/">https://www.a-devgroup.com/teleprompter/best-teleprompter-apps/</a></p>
]]></content:encoded></item></channel></rss>