AI video editor: upload raw footage, describe the edit in chat, and the AI automatically syncs cuts to the beat, adds captions and motion graphics, and exports a short video; it can also be called by Claude/ChatGPT as an MCP.
Brown University student Sulan Zhang (a painting/vlog creator with 65K Instagram followers, @sulansart) and classmate Caleb Pong, who has 7 years of editing experience, started building an AI editor that "edits your own real footage by chat" in the winter of 2025. They first let campus friends try it, then from 2026 used Sulan's own Instagram Reels for a cold start: videos end with "comment 'video' and I'll DM you the link", and the site says three videos she edited with Vyra got 2.9M, 8.7M and 2.1M views. In 2026-05 they launched an MCP connecting to Claude/ChatGPT to ride the Agent hype, three months ago Buffer's review ranked it first, and in autumn it entered YC F26. They also scaled with the paid "Create for us" creator program, short videos with personal tracking links (/go/caleb-pong), and 255 SEO landing pages; the site claims 89K creators.One-line growth story, compiled by us from public sources; speculation is marked separately below.
The main engine is a creator bringing followers: founder Sulan is an Instagram creator with 65,000 followers, the Reels she cuts with Vyra are the product demo, and they end by telling people to comment for a link.
The turbo boost was MCP and YC: in 2026 it launched an MCP connecting to Claude and ChatGPT and joined YC F26, bringing a wave of early adopters from the AI crowd.
By estimate, about 40% of visits come from seeing Reels, and visits are estimated; about 8 in every 100 visits click through to the checkout page.
Model: Lenny's racing growth framework. Engine = a creator bringing followers (the founder cuts her own Reels with Vyra, the videos are themselves the demo and end by telling people to comment for the link; then paid UGC creators and tracking links scale it); turbo boost = launching MCP, joining YC and the Buffer comparison; lubricant = 255 SEO landing pages and old users coming back to cut new videos.
Real data: The biggest channel organic search at 41.77%, top-three channel ranking (organic search, direct, referrals), -7.29% month over month, organic search 100% of search, 46 search terms in total, top search terms (vyra, vyra ai, usevyra, vyra editor, usevyra ai), social sources (X, LinkedIn, Instagram), countries (India 36.15%, US 31.2%, Indonesia 5.69%, Canada 4.25%, Australia 2.93%), bounce rate 38.92%, 3.42 pages per visit, average time on site 1 minute 25 seconds, global rank #312060, all from the Similarweb public page for last month; the public page does not give the shares of the second and third channels. 8,596 visits (about 8,600) to the Stripe checkout page come from quietmoney checkout page data. The dossier's old channel numbers (direct 43.7%, search 41.8%) are an earlier basis and are not used here.
Supporting evidence: Brown University student Sulan Zhang and Caleb Pong, who had edited video for 7 years, started building it in the winter of 2025, first letting campus friends try it; Sulan is an Instagram creator with 65,000 followers who makes drawing and vlog content, the homepage shows 3 Reels she cut with Vyra (2.9M, 8.7M and 2.1M views), and the Reels end by telling people to "comment video and I will DM you the link"; the site has a paid "Create for us" creator program, and Caleb Pong's Facebook Reels use the /go/caleb-pong tracking link; in 2026 it launched an MCP connecting to Claude and ChatGPT, the site says it ranked first in a Buffer comparison, and then it joined YC F26; the sitemap has 255 SEO landing pages; the site claims 89,882 creators, and the iOS app is live.
Inference: The 108,500 total visits is an estimate: the Similarweb public page does not give this site's visits, so it was fitted from the curve of 437 sites with real visit data using global rank #312060, where half of the sites are within 17% error, which is why the channel circles are marked with an approximate sign. The two channel circles are cut at 41.77% and 58.23%, and how each route is split is an inference: searching vyra 45,300 = searching the name after seeing Reels 18,000 + searching editing tools and landing on a page 12,000 + old users searching the brand name 10,000 + MCP and YC 5,300; other channels 63,200 = old users opening directly 30,000 + commenting for the link after seeing Reels or clicking the profile link 24,000 + MCP and YC 9,200. Seeing Reels at 3% visiting, searching editing tools and seeing a landing page at 5% clicking through, and seeing the MCP demo and the YC page at 5% visiting are guesses.
Checkout: checkout.stripe.com
“"Made with Vyra" section: "Real content from real creators. From raw footage to published in minutes."; "Free to start"; iOS App Store link; Discord in footer”Quote from the site
New-user source: SEO (Non-brand search). Direct 43.7%, search 41.8%, referral 8.7%; brand keywords 68.7%, non-brand search about 16.7% (things like "vyra ai" and "vyra editor"); social sources are X, LinkedIn and Instagram, plus an iOS app and Discord; the homepage shows "Made with Vyra" works. Returning users come back via the app and direct visits.
Top channel measured by Similarweb, others estimated by rank
| Attributable new users, last month | Share |
|---|---|
| SEO | 17% |
| Other | 83% |
Direct + brand search 69%. Direct and brand search count as viral only when the site shows sharing mechanics; the largest share wins, under 5% means Other.
| Country | Share of visits |
|---|---|
| India | 36.15% |
| United States | 31.2% |
| Indonesia | 5.69% |
| Canada | 4.25% |
| Australia | 2.93% |
English
X (Twitter), LinkedIn, Instagram
46 keywords · brand ~69% · organic 100% / paid 0%
| Similarweb top 5 keywords |
|---|
| vyra |
| vyra ai |
| usevyra |
| vyra editor |
| usevyra ai |
Non-brand queries: vyra ai, vyra editor
| Category | Share of referrals |
|---|---|
| Business Services | 32.09% |
| Computers Electronics and Technology - Other | 1.33% |
About 8,596 last month (Similarweb estimate), #387 of 606; the month before was 0. Opening checkout is not paying, so this is not revenue.
New-user source: SEO (Non-brand search). New users mostly come from need-based searches Direct 43.7%, search 41.8%, referral 8.7%; brand keywords 68.7%, non-brand search about 16.7% (things like "vyra ai" and "vyra editor"); social sources are X, LinkedIn and Instagram, plus an iOS app and Discord; the homepage shows "Made with Vyra" works. Returning users come back via the app and direct visits.
Free to start; MCP Starter $24/month, MCP Pro $65/month; Vyra AI Starter $54/month, Pro $129/month
Dossier confidence: medium. Parts 1 and 2 come from reading the site and Similarweb data; part 3 from public reporting, founder posts, Wayback snapshots, WHOIS and similar, with speculation marked.
The number: checkout visits is Similarweb's estimate of visits from the site to a checkout payment page; opening checkout is not paying, so this is not revenue. Small sites land on fixed buckets.
Data through last month. Page updated 2026-10-31.