FastPix, Inc.

𝗪𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝗮𝗱𝗱𝗶𝗻𝗴 𝘃𝗶𝗱𝗲𝗼 𝗱𝗼? Spotify added video. Storylane added video. Many creator platforms added video this year. The announcements all rhyme. But the P&L impact? We rarely get to see it. That's what makes 𝗔𝘂𝗱𝗶𝗼𝗕𝗼𝗼𝗺, the numbers, and the trajectory interesting. In 2024 they announced a company-wide video strategy. Not a feature. An initiative. The "we are re-platforming to a shift already underway" kind. Consumers had already moved 75% of some shows' consumption to YouTube. AudioBoom decided to catch the money. Then they showed their work, feature by feature, in the order a product roadmap actually evolves: 🆕 𝗙𝗶𝗿𝘀𝘁, 𝘃𝗶𝗱𝗲𝗼 𝗵𝗮𝗱 𝘁𝗼 𝘄𝗼𝗿𝗸: Ingest, hosting, playback. Video as a first-class asset alongside audio, not a side stream. Not glamorous, but → By June 2025, video was 𝟐𝟎% of platform consumption. 📈 M𝗮𝗱𝗲 𝘀𝗵𝗶𝗽𝗽𝗶𝗻𝗴 𝘃𝗶𝗱𝗲𝗼 𝗲𝗮𝘀𝘆 𝗳𝗼𝗿 𝗰𝗿𝗲𝗮𝘁𝗼𝗿𝘀: Publishing tooling, YouTube distribution, video-ready workflows so producing the video version stopped being extra work → 60%+ of AudioBoom creators shipping video versions of their shows → 𝟐𝟐𝐌+ monthly video views by June 2025. 💸𝗧𝗵𝗲𝗻 𝘁𝗵𝗲𝘆 𝗺𝗮𝗱𝗲 𝘁𝗵𝗲 𝘃𝗶𝗱𝗲𝗼 𝗽𝗮𝘆: Baked-in live-read ads that monetize the show. By H1 2025, video revenue had doubled YoY and hit >𝟏𝟑% of total business. 🔥𝗧𝗵𝗲𝗻 𝘁𝗵𝗲𝘆 𝗺𝗮𝗱𝗲 𝗶𝘁 𝘀𝗰𝗮𝗹𝗲: Showcase, a programmatic marketplace built specifically for video ad inventory, grew +𝟔𝟑% YoY. Spotify and Apple partnerships wired in for H2 2026. → By H1 2026, company-wide revenue up 𝟑𝟎%, adjusted EBITDA up 𝟖𝟎%. Each feature earned its own line on the P&L. If you're a creator platform still deciding whether to add video, the honest question isn't "will my users watch it." Your data probably already answers that. The honest question is 𝐰𝐡𝐢𝐜𝐡 𝐟𝐞𝐚𝐭𝐮𝐫𝐞𝐬 𝐲𝐨𝐮 𝐛𝐮𝐢𝐥𝐝, and in what order, so the business result actually shows up. 𝘛𝘩𝘢𝘵'𝘴 𝘸𝘩𝘢𝘵 𝘢𝘥𝘥𝘪𝘯𝘨 𝘷𝘪𝘥𝘦𝘰 𝘢𝘤𝘵𝘶𝘢𝘭𝘭𝘺 𝘥𝘰𝘦𝘴, 𝘸𝘩𝘦𝘯 𝘺𝘰𝘶 𝘣𝘶𝘪𝘭𝘥 𝘵𝘩𝘦 𝘳𝘪𝘨𝘩𝘵 𝘱𝘢𝘳𝘵𝘴 𝘪𝘯 𝘵𝘩𝘦 𝘳𝘪𝘨𝘩𝘵 𝘰𝘳𝘥𝘦𝘳. ----------------------------- P.S. If your team is on this path, those building blocks above are what we ship at FastPix, Inc.: video ingest, cross-surface player, YouTube-ready publishing, baked-in ad markers, programmatic video delivery, per-clip AI + analytics. In sprints, not seasons 💪

Modern Video, now in Moodle Moodle? The world’s largest open source LMS that hit 500 million learners Modern? Well until now, on its own, Moodle stores a video as a file. It doesn't transcode, caption, add DRM, or measure it. Things every viewer now expects 𝘱𝘢𝘳 𝘥𝘦𝘧𝘢𝘶𝘵. And importantly, a student can finish every default video in your Moodle course without actually watching them. "Complete" just means the student opened the activity. FastPix’s new Moodle plugin brings all the above and fixes what’s missing underneath. Four plugins. Each does one thing: 𝗹𝗼𝗰𝗮𝗹_𝗳𝗮𝘀𝘁𝗽𝗶𝘅: credentials, connection, webhooks, and token signing for private and DRM playback. 𝗺𝗼𝗱_𝗳𝗮𝘀𝘁𝗽𝗶𝘅: the video activity. Uploads stream straight to FastPix, playback is adaptive. 𝗳𝗶𝗹𝘁𝗲𝗿_𝗳𝗮𝘀𝘁𝗽𝗶𝘅: drops a player into any page, summary, or forum post. 𝘁𝗶𝗻𝘆_𝗳𝗮𝘀𝘁𝗽𝗶𝘅: an "Insert FastPix Video" button in the editor. With these you get all the benefits of flawless video streaming FastPix is known for. And our 𝘮𝘰𝘥_𝘧𝘢𝘴𝘵𝘱𝘪𝘹 plugs the thing Moodle can't on its own: measure actual watching. Watching? The unique seconds actually watched, deduped, with seeking ruled out. The grade follows from coverage plus skipping, so a full watch and a skim stop scoring the same. It writes back through Moodle's own completion and grade APIs. 𝗩𝗶𝗱𝗲𝗼 𝘁𝗵𝗮𝘁 𝗱𝗼𝗲𝘀𝗻'𝘁 𝗹𝗶𝗲 𝘁𝗼 𝗠𝗼𝗼𝗱𝗹𝗲. Free and open source, on the Moodle Plugins Directory now. Code: https://lnkd.in/dabNXXVm

📺 Newsrooms aren't losing the audience to better journalism. They're losing it to better experiences and better pipelines. And that's not me. That's the 15th Reuters Institute Digital News Report. Abt 100,000 people, 48 markets. The takes: 📩 Social media and video platforms have overtaken news websites as the main way people reach news 📩 77% watch online news video every week, mostly on third-party platforms. 📩 27% now get news from creators. Not because they trust them, but because they're more entertaining and easier to get. It's what it is. The gut punch for publishers: the audience didn't leave for a better story. They left for a better-𝘥𝘦𝘭𝘪𝘷𝘦𝘳𝘦𝘥 one. Social, vertical, clipped, playing in <1s on a phone. And "better-delivered" is an engineering problem wearing a nice press badge. It's the plumbing between the field and the feed, and the breaking-news window is about 20 minutes: 💻 Monitor the world against your audience + editorial rules  📝 Score > write > deliver ⚡ Pull the upload off a reporter's phone, instantly ☄️ Encode + auto-caption ✂️ Cut that clip while it's still hot, in every aspect ratio, for every platform Aadhan, a short-news app that climbed to the top tier on Google Play, runs its video on FastPix, Inc.. They skipped the part most newsrooms suffer through. They put their engineers on a modern news experience that plays exactly to how audience consumes today. Building in news video? DM me. Always up for comparing notes with people closer to the wires or fires than I am. Every newsroom thinks it's in the business of knowing first. Agreed. But the survivors will also be in the business of arriving first and better 🙌

We don't get to choose what folks like 𝗢𝗥 not like about us. The feature my newest creator-platform customer loves isn't the one we spent the most engineering on. 𝗪𝗵𝗮𝘁 𝗱𝗼 𝘁𝗵𝗲𝘆 𝗹𝗶𝗸𝗲? 𝗔𝗜 𝗰𝗹𝗶𝗽𝗽𝗶𝗻𝗴 𝗔𝗣𝗜 A creator drops in a one-to-two-hour recording, walks away, comes back to 𝟏𝟎+ clips ready for any social platform. The agent picked the viral moments. Top "clippers" - creators doing this at scale - make $5k–$30k a month. For a creator platform, that's not a content trend. It's the feature power users demand, and the reason they stay or leave. It should have been obvious. 𝗕𝘂𝘁 𝗺𝘆 𝗺𝗶𝗻𝗱 𝘄𝗮𝘀 𝗼𝗻 𝘁𝗵𝗲 𝗼𝘁𝗵𝗲𝗿 𝗹𝗮𝘆𝗲𝗿 - uploads that land from any device, players that don't stutter, costs that don't surprise, data that quietly feeds the next recommendation. That layer took years. From upload to first play we're 𝟐.𝟖𝐱 faster than most competitors, and that number is the reason the next part works at all. But AI clipping API is what drove delight & traction. Because creativity is hard enough without the clipping-and-social-platform-rules drudgery. Customers rightfully care about what drives their goals. The hardest call when building products is which voice to listen to next sprint. 🧱 Your own that says the foundation needs more work. ✨ Theirs that names what makes them stay. 𝘉𝘰𝘵𝘩 𝘢𝘳𝘦 𝘳𝘪𝘨𝘩𝘵. 𝘕𝘦𝘪𝘵𝘩𝘦𝘳 𝘪𝘴 𝘦𝘯𝘰𝘶𝘨𝘩 𝘰𝘯 𝘪𝘵𝘴 𝘰𝘸𝘯. So we keep building both layers, and weight the next sprint the best we can. The trade-off never really goes away.  

Vertical streaming just went from experiment to table stakes for OTT. In the last 30 days: → Netflix shipped Clips → Prime Video rolled out scrollable short-form → Peacock partnered with ReelShort on micro-drama originals → Cellcom launched vertical in Israel → Watch IT launched in Egypt → Goldie Behl bet on vertical dramas at TCH x VAM Coverage will treat each as a separate story. The pattern says otherwise. • • • We've been on a lot of OTT calls in the last month. One question keeps coming up — and it's not the one we expected. It's not "should we ship vertical." That's already decided. Every platform we're talking to has it on the roadmap. The real question: "How do we ship this without breaking the long-form app that's paying our bills today?" That's a different problem than the one the market is solving for. And it's the question most of the coverage is missing. • • • The industry is repricing where attention lives. → Media houses are putting TikTok and Instagram creators on payroll → Production budgets are shifting to short-form OTT platforms have two moves: 1. Partner with someone who already cracked vertical (the MX Player x Amazon route) 2. Build the pipeline in-house without disturbing the stack generating revenue today Most pick option two — and underestimate it. • • • Vertical isn't a portrait-rendered version of long-form. It's a different stack underneath. Honest gut-check on where yours stands: 1. Does your player render full-screen mobile-first, or landscape-first with PiP baked in? 2. Does your ingest accept 60-second phone-shot and AI-generated clips at volume — or was it built for 20-minute episodes on a long-form transcoding ladder? 3. Does your player preload the next 10 videos in the scroll, or does each tap trigger a fresh fetch? 4. Do your analytics track 90-second completion, or 25-minute retention? 5. Can your team ship all of that in one quarter without freezing the long-form roadmap? If more than two need work, this isn't a feature anymore. It's a rebuild. • • • The rebuild is what's actually scaring the teams calling us. Not the vertical part. The "without breaking what works today" part. Here's the trap nobody's pricing in: While you spend two quarters rebuilding, early movers spend those quarters on content and audience. By the time you ship, they have a library, a retention curve, and a habit loop. The infra gap closes. The audience gap doesn't. • • • This is what we've been heads-down on at FastPix — vertical and long-form running off the same pipeline. No two codebases. No frozen roadmap. It's also why we keep getting these calls. If you'd rather spend the next two quarters on audience instead of infra: → fastpix.com/contact-us

Douyin shipped 𝟓𝟎,𝟎𝟎𝟎 𝐀𝐈-𝐧𝐚𝐭𝐢𝐯𝐞 𝐦𝐢𝐜𝐫𝐨𝐝𝐫𝐚𝐦𝐚 𝐭𝐢𝐭𝐥𝐞𝐬 𝐢𝐧 𝐌𝐚𝐫𝐜𝐡 𝟐𝟎𝟐𝟔 alone — more than Netflix's entire 28-year library, in one month. The compute-per-second-of-generated-video curve has been collapsing for four quarters straight -> AI video inference just got 10-20× faster & production cost dropped to roughly 1/5 of live-action ByteDance's 𝐒𝐞𝐞𝐝𝐚𝐧𝐜𝐞 𝟐.𝟎, Kuaishou's 𝐊𝐥𝐢𝐧𝐠 𝟑.𝟎, and Shengshu's 𝐕𝐢𝐝𝐮 are doing the lift. Generation isn't the bottleneck anymore. 𝘋𝘦𝘭𝘪𝘷𝘦𝘳𝘺 𝘪𝘴. But most teams underestimate what "AI-generated content" does to the video 𝘥𝘦𝘭𝘪𝘷𝘦𝘳𝘺 pipeline by an order of magnitude. Here's what 50,000 new titles per month actually breaks: ⚜️ 𝐈𝐧𝐠𝐞𝐬𝐭 𝐟𝐚𝐧-𝐨𝐮𝐭. A 30-second microdrama master at 9:16 4K is ~80 MB. 50,000 of those is 𝟒 𝐓𝐁/𝐝𝐚𝐲 of new master content per platform — before transcoding. Direct uploads need to be resumable, multi-part, and CDN-edge-accelerated; a flat S3 PUT pipeline will eat your TTFB for breakfast. ⚜️ 𝐏𝐞𝐫-𝐭𝐢𝐭𝐥𝐞 𝐞𝐧𝐜𝐨𝐝𝐢𝐧𝐠 𝐥𝐚𝐝𝐝𝐞𝐫𝐬. Vertical 9:16 needs its own ABR ladder — phone-first connections sit between 1.5 and 6 Mbps with frequent drops. AV1 + HEVC + H.264 at three rungs each = 9 renditions × 50,000 titles = 𝟒𝟓𝟎,𝟎𝟎𝟎 𝐧𝐞𝐰 𝐞𝐧𝐜𝐨𝐝𝐞 𝐣𝐨𝐛𝐬/𝐦𝐨𝐧𝐭𝐡. Per-title encoding (VMAF-targeted, not flat-bitrate) is the only way the storage bill stays sane. ⚜️ 𝐂𝐨𝐧𝐭𝐞𝐧𝐭-𝐩𝐫𝐨𝐭𝐞𝐜𝐭𝐢𝐨𝐧 𝐚𝐭 𝐯𝐞𝐥𝐨𝐜𝐢𝐭𝐲. AI-generated content is 𝘵𝘩𝘦 class of content that gets re-uploaded to rival platforms within hours. Signed URLs + Widevine/FairPlay + visible/forensic watermarking has to be a build-time concern, not a "we'll add it Q3" concern. ⚜️ 𝐑𝐞𝐜𝐨-𝐞𝐧𝐠𝐢𝐧𝐞 𝐝𝐚𝐭𝐚 𝐢𝐬 𝐧𝐨𝐰 𝐭𝐡𝐞 𝐦𝐨𝐚𝐭. When the catalog is 50× the size of human-written content, 𝘩𝘶𝘮𝘢𝘯 𝘤𝘶𝘳𝘢𝘵𝘪𝘰𝘯 𝘪𝘴 𝘥𝘦𝘢𝘥. Per-user, per-second event telemetry is what trains the model that decides which of the 50,000 titles a viewer sees in their next swipe. If your data layer can't answer "who watched ≥50% of which clip in the last 60 s?" in single-digit ms, your reco engine is 50× slower than the pipeline that fed it. The platforms that survive the AI-content wave aren't the ones with the best generator. They're the ones whose 𝐝𝐞𝐥𝐢𝐯𝐞𝐫𝐲 + 𝐭𝐞𝐥𝐞𝐦𝐞𝐭𝐫𝐲 + 𝐩𝐫𝐨𝐭𝐞𝐜𝐭𝐢𝐨𝐧 𝐥𝐚𝐲𝐞𝐫 can absorb a 50× catalog blow-up without rewriting the stack. Generation is becoming a commodity input. Distribution is becoming the moat. 🎯 If your video backend had to handle 10× catalog growth in 90 days, would you rewrite the encoder, the DRM layer, or the data plane first? Or we can just handle it for ya 😎

Most "AI video" coverage misses the second-hardest problem. In the last week alone:   ⚜️ 𝗙𝗮𝗹.𝗮𝗶 launched HappyHorse-1.0, currently the #1-ranked AI video model, as an official API partner.   ⚜️ 𝗙𝗿𝗲𝗲𝗽𝗶𝗸's video-AI pivot quietly built a $230M ARR business with no VC funding (Fortune, April 28).  ⚜️𝗚𝗼𝗼𝗴𝗹𝗲 started testing an AI chatbot search for YouTube videos. Every one of those headlines is about 𝘨𝘦𝘯𝘦𝘳𝘢𝘵𝘪𝘰𝘯 or 𝘥𝘪𝘴𝘤𝘰𝘷𝘦𝘳𝘺. Almost none of them are about the boring middle:  👁️🗨️ How do you actually deliver an AI-generated 4K vertical clip to a user's phone in under 500ms, transcoded at 5 bitrates?  👁️🗨️ With auto-captions, NSFW-filtered, search-indexed, entitlement-gated?  👁️🗨️ And metered for the creator's monetization split? What the AI-video deliverability stack actually has to do, post-generation?  1. 𝐑𝐞-𝐞𝐧𝐜𝐨𝐝𝐞 𝐟𝐨𝐫 𝐝𝐞𝐥𝐢𝐯𝐞𝐫𝐲. Generation models output mezzanine-quality MP4. ABR ladder, codec choice (HEVC vs AV1), per-title encoding - ” none of that is solved by the generator.  2. 𝐂𝐚𝐩𝐭𝐢𝐨𝐧 + 𝐢𝐧𝐝𝐞𝐱. Speech-to-text on the audio, optional shot-level object/logo detection, then push the metadata into a search index. AI search is only as good as the upstream extraction.  3. 𝐌𝐨𝐝𝐞𝐫𝐚𝐭𝐢𝐨𝐧. NSFW detection, brand-safety classification, copyright/likeness checks. Synchronous or your moderation queue floods within hours.  4. 𝐄𝐧𝐭𝐢𝐭𝐥𝐞𝐦𝐞𝐧𝐭 + 𝐦𝐨𝐧𝐞𝐭𝐢𝐳𝐚𝐭𝐢𝐨𝐧. Creator gets X% of view-revenue. Per-asset signed playback. Webhooks into the billing system on every state change.  5. 𝐓𝐞𝐥𝐞𝐦𝐞𝐭𝐫𝐲. View-start, completion %, buffer ratio, sliced per-creator. The data is the moat for the recommendation engine that drives the next generation cycle. Everyone loves the magician. No one studies the trapdoor beneath the stage. But that’s where the real work happens. That’s where FastPix operates. Take a closer look --> www.fastpix.io 

Trends change, your body changes… your wardrobe? In the past year, our platform grew, our features grew…so it was about time to refresh our dashboard UI. Here is how we went about it:   - Most-used actions are now front and centre - Stream health, encoding status, and playback data are visible without hunting - API keys and credentials are cleaner to manage - Navigation rebuilt around what developers actually do, not how we built the product   We didn't redesign for aesthetics. We redesigned so developers stop losing time inside our UI and get back to building.   It's a START. We have a few fixes, a mood board refresh, and a larger website refresh in the works. We've got miles to go before we sleep, but it's ok to pause, reflect, and share, right?   Sign in and take it for a spin, we'd love to know what you think 👇 https://lnkd.in/gHJX9JPK