Skip to content

AI girlfriend site profile · Updated October 7, 2026

FantasyGF logo

FantasyGF

An AI girlfriend site: what it offers, what people say, and the numbers.

I bought a FantasyGF membership and tested it myself. Then I built a data-tracked profile of fantasygf.com, refreshed monthly. Data through September 2026, from DataForSEO, Ahrefs, Tranco and Google Trends.

FantasyGF in numbers

Search visits a month, worldwide, the last twelve months.

877visits, Sep 26

#13
among the sites I track
1.3K
US name searches, Aug 26
#727K
Tranco, all domains
Free
free tier, paid plans on top

I bought a FantasyGF membership and tested it myself, feature by feature. I’m an AI girlfriend enthusiast and a data researcher, and both shape every page on this site: my own test tells me what the product does, but I don’t grade FantasyGF on my feel alone.

I weigh what its users publish, and I track it the way an analyst tracks a market, with a fixed set of indicators that show how big a platform is, whether it is growing or fading, and how it compares with every other AI girlfriend site I follow.

This page answers three questions in order: what you get, what people are saying, and what the numbers show. Part 1 comes from my own paid test and FantasyGF’s feature list and pricing, Part 2 from the videos, Reddit threads and press I read in full, and Parts 3 and 4 from traffic and search data. That is also how I rank the best AI girlfriend sites: tested first-hand, then judged by what users say and by the numbers.

Part 1 · What it offers

What FantasyGF offers

FantasyGF (fantasygf.com) is an adult AI girlfriend platform: you create or pick a companion and talk to it, with voice, AI-generated images on top of text chat. I bought a membership and tested it myself; this is what it offers, and Part 2 is what users say about each part of it.

Features
Free tier with core chat access, Voice call capability, NSFW chat and image generation, Detailed character creation, Cryptocurrency payments
Free tier
The free tier gives one trial character a month, 16 blurred images and limited chat.
Plans
Free
Extras
Free tier available, though it's more of a limited trial than a fully open plan. Paid plans run $12.99/month, roughly $9.99/month billed quarterly, or roughly $5.99/month billed annually (about $69.99/year), against a listed regular price of $25.99/month. Crypto accepted (BTC, ETH, USDT, USDC, LTC, SOL, MATIC). Note: fantasygf.ai now redirects to fantasygf.com.
Operator
Advanced AI Labs S.R.L. (company number 49734069, named in the site footer; the pricing grid is a near copy of DreamGF's, which is why the two get called sister sites, but the footer names a different Romanian company from DreamGF's DreamAI SRL), Oradea, Bihor county, Romania, since 2023 (fantasygf.com was registered in September 2023; the brand ran as FantasyGF.ai first, was listed as a Candy AI alternative by March 2024, and search traffic to the .com begins in November 2025); domain registered September 2023
Best for
Users wanting a feature-rich free-entry NSFW girlfriend experience with voice and image generation
Not for
Users requiring a mobile app or needing cross-session memory on the free tier

Features and prices from the platform’s own pages; operator from press coverage (fantasygf.com home and pricing pages, Domain registration record (DataForSEO WHOIS), Reddit list posts (for the 2024 .ai name), My own paid month). Official site and social accounts are listed at the end of the page.

Part 2 · What people say

What people say about FantasyGF

I bought a membership and tested FantasyGF myself, but I don’t take my own word for it. Many users have a more grounded view of a platform this popular than one reviewer can, and I would rather not give it too much or too little credit from my own intent, so I treat what users publish and ask as the main signal for how the brand, and each of its features, is regarded. On October 7, 2026 I found 3 YouTube videos with 485 views between them, 3 Reddit threads that Google surfaces, and 0 news articles in the last year. I did not stop at the titles: I read 3 video transcripts (2,080 words), 3 threads with their top comments, and classified each one as positive, mixed, neutral or negative.

Favourability at a glance

Each item was classified after I read it, not from its title. A walkthrough or a passing mention counts as neutral.

SourcePos.MixedNeut.Neg.Fav.
YouTube videostranscripts read in full101133%
Reddit threadspost and top comments102033%
All sources203133%

In one line: YouTube videos 33% positive, 1 negative of 3; Reddit threads 33% positive, 0 negative of 3. The gap between YouTube videos on one side and Reddit on the other is the main sentiment finding on this page, and the sections below explain it.

Each feature, and what people think of it

What users say about each part of it, drawn from the videos, threads and articles below, and whether my own paid test agrees.

FeatureVerdictWhat people say
Chat and roleplayMixedReddit's one-line verdict is 'feels more like a virtual dating app' (1 thread). The hands-on video found the chat boring and unrealistic, with lines like 'taste the air after you whisper my name', though replies arrived in about three seconds (1 of the 3 videos). The promotional video promises a companion that learns your preferences (1 video).My test: Partly disagree. 'Boring' undersells it: my companion stayed in character and kept the relationship dynamic across sessions, which is above the category average; the purple prose the reviewer mocks is real, and you can tone it down in the character settings.
MemoryMixedThe pricing page is specific: 110 messages of chat memory on Premium, 300 on Ultra, described as five times larger. The promotional video says she learns your preferences over time (1 of the 3 videos), and the hands-on test never got far enough to find out. No user discusses memory (0 threads).My test: Agree. The numbers are right, with a catch: on the free tier and across days, recall was noticeably thinner than inside one session, so the 110-message window is a rolling one, not a diary.
Image generationMixedTwo videos sell it (casual to fully explicit on request) and the hands-on test found the picture for 'show me your outfit now' was a random snow scene unrelated to the chat (1 video). Reddit lists it among the better adult image generators (1 thread). Pictures cost from 2 tokens on a 100-token Premium allowance.My test: Partly disagree. Images followed my companion's defined look consistently; they follow the conversation less well, which is what he hit. The allowance of roughly fifty pictures a month is generous for a $5.83-a-month annual plan.
VideoMixedA hugging-video tutorial shows the clips take a few clicks (1 video), and the hands-on test found a clip generated in about fifteen seconds that 'just moves' the source image, at 20 coins, a fifth of the monthly allowance (1 video). The site now quotes 5-second videos from 6 tokens.My test: Agree. The kissing and hugging clips are a genuine extra most rivals lack, and they are animations of a still, not scenes; the price has come down since the test if the pricing page is current.
Voice messages and callsNegativeThe hands-on test is damning: voice messages failed with 'developers have been notified', the sample that played was 'absolutely terrible', and two call attempts never connected (1 video). The site lists calls at 8 tokens a minute on Premium and 4 on Ultra, up to 12 minutes a month on the smaller plan.My test: Partly disagree. Calls connected and worked in my month, so the tester hit an outage; the voice itself is ordinary text-to-speech, and twelve minutes a month on Premium is a taste, not a feature.
Characters and customisationPositiveA large roster, a builder that makes a companion 'in 10 seconds', and 5 free characters a month on Premium, 15 on Ultra. The hands-on test liked the modern layout but said several faces 'look very AI' with odd eyes (1 video). Companions you create are public unless you set them private (site).My test: Agree. The builder is quick and the results matched what I set; a few of the stock faces are off, and the public-by-default setting is the one I would change first.
Free tierNegativeThe site defines it: one trial character a month, 16 blurred images and limited chat, with 'basic features only'. No video or thread reviews the free tier (0 videos, 0 threads); the one hands-on tester paid for premium before starting, and the list posts never mention a free plan at all.My test: Agree. It is enough to see the interface and the roster, and the blurred images are a sales pitch rather than a trial; the chat cap arrived within a session for me.
Price and tokensMixedPremium is $25 a month or $69.99 for the first year ($5.83 a month), Ultra $49.99 a month or $299.99 a year, with a three-month option at 60% off. Tokens pay for pictures (from 2), videos (from 6), voice messages (0.5) and calls (8 a minute). The tester paid about $13 and called the video price steep (1 video); Reddit rates it level with Candy on value (1 thread).My test: Partly disagree. The tester is right on monthly value and wrong on annual: $25 a month is poor, $69.99 a year with 100 tokens a month is one of the cheaper ways into a full-featured site. The grid is a near copy of DreamGF's.
Privacy and billingMixedThe promotional video promises 'end-to-end privacy for all communications' (1 video). The site itself says the charge appears under a neutral payment-company name, accepts crypto, and that created companions are public unless made private. No user reports a billing problem (0 threads).My test: Disagree. Chats are stored on the company's servers and companions are public by default; the neutral descriptor and crypto option are the real privacy features, and they are good ones.
Safety and legitimacyMixedThe footer names Advanced AI Labs S.R.L. of Oradea, Romania, with a company number, and the brand has a Crunchbase profile (1 article, unread). It is widely called a sister site of DreamGF because the pricing grid matches, but the footer names a different company. No incident, breach or scam report exists anywhere I looked (0 sources).My test: Agree. It is a legitimate, identifiable operator; I keep its numbers separate from DreamGF's; whatever the business relationship, the community figures belong to different brands.

How I read this

Sentiment is the hardest thing to measure honestly. My own test tells me what the product does; what other people publish tells me how it is regarded, and for a brand this size the second matters more. So for every source I read the full text, not the title: the transcripts of the most-watched videos, Reddit threads with their top comments and upvotes, press articles in full, and the exact words people add after the brand name in search. A title tells you what someone set out to test; only the full text tells you what they found.

The most-watched videos, and what they actually say

For each video I read the transcript, wrote down what it says, and added my own read on it. Most viewed first.

  • Want to Create Realistic Hugging Videos? FantasyGF AI Makes It Easy!AI Girlfriend · 1y ago · 1:16 · 204 views

    Neutral, no verdict · what the video says

    A 76-second walkthrough of the hugging-video generator: pick the AI hug template, upload clear, well-lit photos or choose existing avatars, decide who appears, preview, press generate and download the clip to share. It is narrated in the site's own voice and ends by pitching the feature as a tool for emotional connection when physical hugs are not possible. There is no test, no price and no opinion; it exists to show that the feature is there and takes a few clicks.

    My read: A feature advert, not a review, and I treat it as one. I agree the tool is this simple to use; what the video leaves out is that each clip draws on the monthly token allowance, so the 'start creating today' line costs more than it sounds. Useful only as proof that video shipped in 2025.

  • FantasyGF Review 2026 | My Girlfriend is AI...Best AI Reviews · 1mo ago · 5:42 · 189 views

    Negative · what the video says

    A paid, hands-on test by the Heaven Girlfriend channel, September 2026. The reviewer buys premium for about $13, likes the modern layout and large roster but says several faces 'look very AI'. Then almost everything fails: the voice message will not play, the sample that does is 'absolutely terrible', two calls at three tokens a minute never connect, the chat is called boring ('taste the air right after you whisper my name'), and the image for 'show me your outfit now' is a random snow scene. Video at 20 coins, a fifth of the monthly allowance, works but only animates that image. Rated low, with a promise to retest.

    My read: The only honest test on record, and I agree with about half of it. The chat in my month was warmer than 'boring' and stayed in character, and voice calls connected for me, so I partly disagree there; the video price, the random image and the flat text-to-speech voice I recognise. Reply speed of about three seconds matches my test.

  • FantasyGF.AI Review: Can AI Love Be Real? Features, Pricing & Real User Experience!AI Girlfriend Mojo · 1y ago · 3:49 · 92 views

    Positive · what the video says

    A scripted feature tour with no hands-on use. It lists six selling points: a companion that learns your preferences over time, 'emotional intelligence' that senses your mood, images from casual to fully explicit on request, creative scenarios, 'end-to-end privacy for all communications including NSFW content', and support 'just an email away'. Every claim is stated rather than shown, the voice-over reads like a brochure, and the video closes by inviting viewers to sign up. Ninety-two views in a year.

    My read: A brochure read aloud, and the privacy line is the part I disagree with. The site's own pricing page says companions you create are public unless you make them private, and chats sit on its servers; 'end-to-end' is a term of art the video borrows without earning. The image and scenario claims are fair; the support claim was the weak spot in my test.

What the videos add up to

Reading across the 3 transcripts, this is the picture that emerges.

Of the 3 most-watched videos whose transcripts I read, 1 is positive, 1 is a neutral walkthrough with no verdict and 1 is negative.

Three videos, 485 views between them, and only one is a test. The Heaven Girlfriend review (September 2026) paid for premium and found voice messages failing, calls not connecting, a flat text-to-speech voice, a chat it called boring and an image that ignored the prompt. The other two, a hugging-video tutorial and a feature tour, are scripted promotions with no hands-on use at all.

What the promotional videos claim is the feature list the site sells: images from casual to explicit, short video clips, scenarios, a companion that learns you, and 'end-to-end privacy'. What the test found is that the features exist and the execution wobbles, with video at 20 coins, a fifth of the monthly allowance, as the one thing that worked as advertised.

The gap between the two is the story of this brand on YouTube. Nobody independent has praised it; the praise is bought or self-published, and the single outside tester came away unimpressed while promising to retest. I weight the test, note that it ran during what looks like an outage of the voice system, and use my own month to split the difference.

What gets praised

  • Images and short video clips exist and are easy to generate (2 of the 3 videos, both promotional)
  • Modern layout with a large roster of companions (1 video, the hands-on test)
  • Replies arrive in about three seconds (1 video)
  • A companion that learns preferences and runs creative scenarios (1 video, promotional)

What gets criticised

  • Voice messages failed to play and calls would not connect during the test (1 video)
  • The voice that did play was flat text-to-speech with no character accent (1 video)
  • Chat called boring and unrealistic, 'nobody texts like that' (1 video)
  • Image generation ignored the prompt and returned a random scene (1 video)
  • Video costs 20 coins a clip, a fifth of the monthly allowance (1 video)

What Reddit says, thread by thread

The threads Google surfaces for "fantasygf" on Reddit, read with their top comments. Upvotes and comment counts as they stood on October 8, 2026; most upvoted first. Screenshots show the post as it is publicly visible.

What Reddit adds up to

Reading across the 3 threads, weighted by upvotes.

Of the 3 threads I read, 1 is positive, 2 are neutral and none is negative.

Three threads name the brand and all three are list posts. A top-ten ranking gives it 4.5 out of 5 and calls it a virtual dating app with voice, image and video; two older lists put it among Candy AI alternatives and among adult image generators. Not one comment in any of them comes from someone who has used FantasyGF, and the most active thread is colonised by three accounts pushing a rival.

The praise is the feature list and the position. Voice, images and video in one product, at a price level with Candy's, is why it keeps appearing on these lists, and the 'virtual dating app' line is accurate. The criticism is absence: no complaint, no billing thread, no memory thread, which for a brand with 1,300 monthly searches tells me its users do not gather on Reddit.

Seven of the ten threads the search returned were noise: a volleyball game, astrology, FIFA Mobile, a streamer, and three chatbot threads where the brand is not named in any text I could read. I dropped them rather than attribute them.

What gets praised

  • Rated 4.5 out of 5, level with Candy AI, as a virtual dating app with voice and image/video (1 thread)
  • Listed as a Candy AI alternative since March 2024 (1 thread)
  • Listed among the better adult image and video generators of 2025 (1 thread)

What gets criticised

  • No first-hand user account in any thread (3 threads)
  • The most-upvoted thread is padded with near-identical praise for a rival from three accounts (1 thread)
  • Framed as a porn generator on a list where it does not belong (1 thread)

In the press

The latest articles that mention FantasyGF, newest first. For each one: what it says, then my take on its view of the brand.

What people type into Google before they sign up

Google's own suggestions after "fantasygf" and the related searches Google Trends ranks highest over the last 12 months, retrieved October 7, 2026.

These are the questions people have before they sign up, in their own words. I group them by what they tell me, because the mix says more than any single search: a brand searched mostly with “free” has a different audience from one searched with “review”.

ThemeSearchesWhat it tells me
Other"fantasygf ai"Related searches that do not fit a theme.

Rising alongside it, last 12 months

  • “myanima”Breakout
  • “herahaven”an AI girlfriend site I track · its page+500%
  • “nectar ai”an AI girlfriend site I track · its page+100%
  • “ai girlfriend”an AI girlfriend site I track+90%

Google Trends “rising” related queries; “Breakout” means growth over 5,000%.

My take

People search for the old name. The branded keyword the site ranks first for is 'fantasygf ai', 590 searches a month, while the site lives at fantasygf.com; the top Trends query is the same. That is a year of marketing under the .ai name still paying off, and it is why the .com's traffic series starts from zero in November 2025 even though the brand is older. There is no autocomplete for the brand, which is unusual for 1,300 monthly searches and tells me Google has not settled on what people want next.

The rising queries are rivals, and they are rising fast. Google Trends shows 'myanima' as a breakout, 'herahaven' up 500%, 'nectar ai' up 100% and 'ai girlfriend' itself up 90% next to the brand, with 'candy ai' among the top related terms. People who search for FantasyGF are comparison shopping across the mid-market, which fits the Reddit record of list posts. The $4.04 cost per click says at least one rival is paying to appear on the brand's name.

What the search data cannot see is the point of the product. The site ranks 80th for 'ai girlfriend' and 22nd for 'ai dominatrix', so search is a side door; the Tranco rank, which samples all traffic, improved from about 1,103,000 to 727,000 in five weeks, so visitors are coming from somewhere else, most likely ads and the affiliate lists that make up its Reddit presence.

YouTube videos and their transcripts, news and search suggestions via SerpApi; Google Trends via SerpApi; Reddit threads read directly from reddit.com in a browser. Retrieved October 7, 2026. Transcripts, threads and articles read and classified on October 8, 2026.

Part 3 · Traffic rank and reach

How much traffic FantasyGF gets, and where it ranks

How I track this: search-traffic estimates miss direct visits, apps and social, so for all-traffic reach I use Tranco, a daily ranking of every domain on the internet built from DNS and browser data (Cloudflare Radar, Cisco Umbrella, Majestic, Farsight and Chrome), and I check it against Google’s own Chrome UX Report each month. I record the rank every day and compare it only with the the AI girlfriend sites I track, because that is the list people actually choose from.

On the Tranco list of October 6, 2026, fantasygf.com ranked #726,988 among every domain worldwide, up 376,433 places since August 31, 2026. Among the AI girlfriend sites I track, that makes it #13, behind GirlFriendly.ai and GirlfriendGPT.

How I read this

The number that matters most in this category is not the worldwide rank but the position among AI girlfriend sites. I read the daily chart for direction: a line climbing toward the top means the platform is gaining ground on everyone else, and a flat line at a high rank means a settled audience. Tranco is a rank, not a visit count, so I never convert it into visits.
#13
among the AI girlfriend sites I track, on the combined data
#726,988
all domains worldwide, Tranco October 6, 2026
Not in top 1M
Chrome UX Report tier, September 2026

Tranco rank, day by day. Higher on the chart is better; hover or tap a point for the exact rank.

700K825K950K1.1M1.2MAug 31Sep 5Sep 10Sep 15Sep 20Sep 25Sep 30Oct 6727K
Low 727KHigh 1.1MAverage 956K
Hover, tap, or use ← → to inspect a day

AI girlfriend sites, ranked

The top 10 of the sites I track: AIGirlfriends.ai first as the editor's pick after testing, then ordered by Google US position for "ai girlfriend" (DataForSEO, September 2026), then by Tranco all-traffic rank for sites that do not rank for it; this platform highlighted.

  1. 1AIGirlfriends.ai#1#633,137
  2. 2RomanticAI#2#411,319
  3. 3Kupid.ai#7#254,630
  4. 4Nomi.ai#9#131,225
  5. 5Candy AI#11#12,246
  6. 6Couple.me#15#985,985
  7. 7Luvr.ai#39#2,762,931
  8. 8DreamGF#53#287,472
  9. 9LoveCoreAI#55#2,343,466
  10. 10Fantasy.ai#62#588,409
  11. 11GirlFriendly.ai#75#1,073,902
  12. 12GirlfriendGPT#79#33,020
  13. 13FantasyGF#80#726,988
  14. 14CrushOn AI–#5,840
  15. 15SpicyChat–#6,952

See all the sites we track →

Its positions on the category searches

fantasygf.com's Google positions on the category's searches. DataForSEO, United States, September 2026.

Lists of shared keywords mix in generic terms, so the comparison I trust is the category’s own searches: the people typing them have not chosen a brand yet. In September 2026, fantasygf.com held position #80 for “ai girlfriend” in the US.

SearchUS searches / monthPosition
“ai girlfriend”135,000#80
“chat girlfriend”22,200#84
“ai gf”22,200#87
“ai girlfriend chat”14,800#99
“ai girlfriend generator”14,800#99
“ai girlfriend free”14,800#91
“girlfriend ai”12,100#98

How I read it. 66% of this site’s top-keyword traffic comes from people typing its own name, so the category searches are what bring in people who have never heard of it. Position #80 for "ai girlfriend" is off page one, so for now the site depends on its name rather than on the category term.

Category positions, United States, DataForSEO. Source: DataForSEO

Tranco blends Cloudflare Radar, Cisco Umbrella, Majestic, Farsight and Chrome data into one daily ranking of the top million domains. Chrome UX tiers via the Chrome UX Report (CC BY 4.0). Retrieved October 7, 2026.

See where every AI girlfriend site stands

Every platform I track, ranked by the same indicators, updated monthly.

Part 4 · The numbers

The numbers, all in one place

Every indicator I track for FantasyGF: the latest value, how it changed, and the period it covers. Retrieved October 7, 2026. Sources are under the table, and the story the numbers tell comes right after it.

How I read this

These are the indicators I track for every AI girlfriend site, in the same order, so the pages can be compared side by side. Traffic tells me size, brand searches tell me demand, the popularity rank tells me standing against the rest of the category, and links tell me staying power. I always read the trend before the level: a platform at 500K visits and falling is a different story from one at 300K and doubling.
MetricLatestChange
Search traffic
Organic visits, worldwideSeptember 2026877▲ +13.5% vs Aug 26
Same month a year earlierSeptember 20251▲ +87600.0% since then
Organic visits, United StatesSeptember 202660068% of worldwide
Last 12 months, totalOct 25 to Sep 264,416—
Peak in 24 monthsSeptember 2026877now at 100% of peak
Low pointSeptember 20251▼ 50% below the Jul 25 peak
Brand demand
"fantasygf" searches, USAugust 20261,300 / mo▲ +30.0% vs Sep 25
"fantasygf" searches, worldwideAugust 20264,400 / mo—
Google Trends interest, worldwideweek of October 4, 202655 / 10055% of the 12-month peak
Google Ads cost per click, USOctober 7, 2026$4.04—
Google position for "ai girlfriend", USSeptember 2026#80135,000 searches / mo
Popularity, all traffic
Rank among the AI girlfriend sites I trackSeptember 2026#13Google US rank for "ai girlfriend" #80
Tranco rank, all domains worldwideOctober 6, 2026#726,988▲ up 376,433 since August 31, 2026
Authority and links
Domain Rating by AhrefsOctober 7, 202651 / 100—
DataForSEO domain rankSeptember 202663 / 100—
Referring domainsSeptember 20261,283▼ -15.3% vs Sep 25
BacklinksSeptember 202642,101▲ +135.9% vs Sep 25
Keywords ranking in GoogleSeptember 2026768 worldwide235 in the US
Community and coverage
YouTube videos about itOctober 7, 20263485 combined views
News articles, last 12 monthsOctober 7, 20260—
Sources read and classifiedOctober 8, 202662 positive · 0 mixed · 1 negative

Sources. Traffic, keywords, links, category positions and WHOIS: DataForSEO. Search volume and cost per click: Google Ads data via DataForSEO. Domain Rating: Ahrefs. Popularity ranks: Tranco and Google’s Chrome UX Report. Trends, videos, news and search suggestions: Google and YouTube via SerpApi. Reddit: read directly. Discord, TikTok and Facebook: each platform's public page. Starting price: the platform’s own pricing page. ▲ and ▼ mark a change of more than 3%.

Reading the data

What the numbers tell me about FantasyGF

First the facts, each one derived from the table above and nothing else. Then the story I read in them.

The facts

FindingDetail
Traffic up 13% on the monthEstimated organic search traffic in September 2026 was 877 visits worldwide, +13.5% versus August 2026 and +87600.0% versus September 2025.
A 50% collapse, then a recovery to 43850% of peakTraffic fell from 2 visits in July 2025 to 1 in September 2025, then climbed back to 43850% of that earlier peak by September 2026.
68% of visits come from the United StatesThe United States accounts for 600 visits, followed by India and United Kingdom.
1.3K people a month search the brand name in the USGoogle Ads data puts US searches for "fantasygf" at about 1,300 a month in August 2026 (+30.0% since September 2025), and worldwide searches at about 4,400.
66% of top-keyword traffic is branded66% of the traffic from the top 10 US keywords comes from people typing the brand name, which indicates a meaningful share of visits from generic, non-branded searches.
1.3K referring domains, Domain Rating 511,283 websites link to fantasygf.com, -15.3% over the last 12 months; its Domain Rating by Ahrefs is 51 and its DataForSEO domain rank is 63 (71 a year ago).
Position #80 on Google for "ai girlfriend"In September 2026, fantasygf.com ranked #80 in the US results for "ai girlfriend", the category's own search term, which is searched about 135,000 times a month.
#13 among the AI girlfriend sites I trackFantasyGF sits #13 among the sites I track on the combined data, ranking #80 in United States Google results for "ai girlfriend", behind GirlFriendly.ai and GirlfriendGPT. Among all domains on the internet it ranks 726,988th in the Tranco list of October 6, 2026, up 376,433 places since August 31, 2026.
One real test found voice and calls broken; the other two videos are the brand's own marketingOf the 3 most-watched videos whose transcripts I read, 1 is positive, 1 is a neutral walkthrough with no verdict and 1 is negative. The one hands-on video found voice and calls broken and the chat flat, while the two promotional videos list features without testing them.

My take: the story in the numbers

What moved, and why I think it moved: the numbers read together, with the reviews in hand.

The 87,600% growth figure is a domain switch, not a boom. fantasygf.com had no search traffic until October 2025, then 199 visits in November, because the brand lived at FantasyGF.ai until then: the year-old videos call it FantasyGF.AI, Reddit listed it as a Candy alternative in March 2024, and the branded keyword it ranks first for is still 'fantasygf ai'. The link data shows the move happening: referring domains to the .com went from 28 in June 2025 to 654 in July and 1,453 in August, backlinks from 289 to 19,117, and the domain rank jumped from 9 to 71 between May and August 2025. That is the .ai site's link equity being redirected to the .com, and the traffic followed three months later.

Since the switch, the line has risen every quarter from a small base. Worldwide visits went from 199 in November 2025 to 311 in April 2026, 477 in July and 877 in September, with 600 of September's from the US. The keywords behind it are category terms at modest positions: 'ai girlfriend' at #80, 'adult ai chatbot' at #34, 'ai dominatrix' at #22. Search is a side door for this site; what matters is that the door keeps opening a little wider.

All-traffic reach is larger than search suggests, and improving. The Tranco rank went from about 1,103,000 at the end of August to 727,000 on October 6, a gain of nearly 400,000 places in five weeks, putting it at #13 of the AI girlfriend sites I track. Brand searches held flat at 1,000 to 1,300 US a month all year, so the Tranco gain is not word of mouth; it is paid and affiliate traffic, which fits a Reddit record made entirely of list posts and a YouTube record that is two parts marketing to one part test.

The link profile churns the way affiliate links do. Referring domains peaked at 1,678 in December 2025, fell to 1,056 by July 2026 and recovered to 1,283 in September, while backlinks swung from 32,960 to 57,085 in August and down to 42,101 a month later. Fifteen thousand links appearing and disappearing inside a month, with the domain count moving a tenth as much, is sitewide placements on a few partner sites being switched on and off. The domain rating of 51 is the one solid thing in it.

What users say explains the ceiling. The one hands-on video in September 2026 found voice messages failing, calls not connecting and chat that read as purple prose; Reddit's only line on the brand is 'feels more like a virtual dating app'. The feature list and the $5.83-a-month annual price are what put it on the comparison lists that drive its traffic; the execution is what keeps the brand searches flat while rival names (myanima, herahaven, nectar ai) break out in Trends next to it.

Verdict. Growing on every indicator I can measure, from a base that only exists because of a domain move, with flat brand demand underneath. If the Tranco gain holds through October it is a real audience and the ranking should rise; if it fades with the ad spend that produced it, the search line of under a thousand visits is what is left. The thing to watch is the brand term: at 1,300 a month it has not moved all year.

Written on October 8, 2026 from the figures above and the sources in Part 2.

Part 5 · Over time

How the numbers moved

One chart per indicator. Hover or tap any point for the exact value and the change against the month before and the same month a year earlier.

Search traffic, worldwide

Visits from Google results each month, 24 months.

The shape is the story: a 50% fall between July 2025 and September 2025, then a climb back to 43850% of the old peak by September 2026. A drop that size usually means Google stopped ranking the site for a while; the recovery means it got back in.

877
latest, Sep 26
877
peak, Sep 26
4.4K
last 12 months
02505007501.0KOct 24Dec 24Feb 25Apr 25Jun 25Aug 25Oct 25Dec 25Feb 26Apr 26Jun 26Sep 26877
Low 0High 877Average 184
Hover, tap, or use ← → to inspect a month

Estimated organic search traffic. Source: DataForSEO

Search traffic, United States

The US is 68% of worldwide visits.

The US line moves with the worldwide one (+59900.0% year on year against +87600.0% worldwide), so whatever is driving the overall trend is not a regional story.

600
latest, Sep 26
02505007501.0KOct 24Dec 24Feb 25Apr 25Jun 25Aug 25Oct 25Dec 25Feb 26Apr 26Jun 26Sep 26600
Low 0High 600Average 130
Hover, tap, or use ← → to inspect a month

Estimated organic search traffic. Source: DataForSEO

Searches for “fantasygf”, United States

Brand searches are the best demand signal I have: nobody types a brand name unless they already know it. Google Ads rounds volumes into bands, so flat stretches are normal.

1.3K
latest, Aug 26
+30.0%
vs Sep 25
$4.04
US cost per click
05001.0K1.5K2.0KSep 25Nov 25Jan 26Mar 26May 26Aug 261.3K
Low 880High 1.3KAverage 1.1K
Hover, tap, or use ← → to inspect a month

Monthly search volume, Google Ads keyword data. Source: DataForSEO

Searches for “fantasygf”, worldwide

The US is about 30% of worldwide brand searches, and worldwide demand is -18.5% year on year. Demand and traffic are rising together, which is the healthiest pattern there is.

4.4K
latest, Aug 26
02.5K5.0K7.5K10.0KSep 25Nov 25Jan 26Mar 26May 26Aug 264.4K
Low 3.6KHigh 5.4KAverage 4.5K
Hover, tap, or use ← → to inspect a month

Monthly search volume, Google Ads keyword data. Source: DataForSEO

Google Trends, last 12 months

Relative interest in "fantasygf" worldwide on a 0 to 100 scale, where 100 is the busiest week. A trend, not a count. Canada is where interest runs highest.

55 / 100
now
August 30
peak week
0255075100Oct 5Nov 16Dec 28Feb 8Mar 22May 3Jun 14Jul 26Sep 6Oct 455
Low 0High 100Average 27
Hover, tap, or use ← → to inspect a week

Google Trends via SerpApi. Source: DataForSEO

Google Trends, last 5 years

Interest only became measurable in December 2023, so this is a young brand by search standards.

0255075100Oct 21May 22Dec 22Jul 23Feb 24Sep 24Apr 25Nov 25Jun 26Oct 2626
Low 0High 61Average 15
Hover, tap, or use ← → to inspect a month

Google Trends via SerpApi, monthly average of weekly values. Source: DataForSEO

Around the brand on Google Trends

Where interest is strongest, and which related searches are growing fastest.

Most interested countries

  1. Canada0
  2. United States0
  3. United Kingdom0
  4. India0
  5. Brazil0
  6. Germany0

Fastest-rising searches

  • myanimaBreakout
  • herahaven+500%
  • nectar ai+100%
  • ai girlfriend+90%

Top related searches

  • fantasygf ai100
  • ai girlfriend16
  • candy ai15
  • herahaven3
  • nectar ai2
  • myanima1

Google Trends via SerpApi, retrieved October 7, 2026. “Breakout” means growth over 5,000%.

Where its visitors come from

Estimated organic visits by country. United States leads with 70% of the total.

The top three countries make up 82% of visits, so this is a concentrated audience; a ranking change in one of them would move the whole number.

CountryVisitsShare
United States61470.0%
India647.3%
United Kingdom384.3%
Canada182.1%
Poland121.4%
Australia70.8%
Sweden60.7%
Philippines50.6%
Germany40.5%
France40.5%
Indonesia40.5%
Turkey40.5%

22 countries checked; together they account for 91% of estimated worldwide visits.

Organic traffic by country. Source: DataForSEO

The searches that bring people in

US Google, ranked by estimated visits. 66% of this traffic is people typing the brand name, so the audience mostly already knows it.

KeywordPos.SearchesVisits
fantasygf aibrand#1590377
ai girlfriend#80135K122
adult ai chatbot#345.4K18
chat girlfriend#8422.2K13
ai girlfriend chat#9914.8K10
ai dominatrix#221.3K7
girlfriend ai#9812.1K6
ai girlfriend generator#9914.8K6
ai girlfriend free#9114.8K6
ai gf#8722.2K4

Ranked keywords, United States. Source: DataForSEO

Websites linking to it

Unique referring domains pointing to fantasygf.com. More linking sites usually means more staying power in Google.

The site is losing linking domains (-15.3% in 12 months). That rarely hurts straight away, but it is the slow leak I would watch.

1,283
referring domains
51 / 100
Domain Rating by Ahrefs
63 / 100
DataForSEO domain rank
05001.0K1.5K2.0KOct 24Dec 24Feb 25Apr 25Jun 25Aug 25Oct 25Dec 25Feb 26Apr 26Jun 26Sep 261.3K
Low 7High 1.7KAverage 842
Hover, tap, or use ← → to inspect a month

Referring domains and domain rank. Source: DataForSEO

Total backlinks

42.1K
latest, Sep 26
025.0K50.0K75.0K100KOct 24Dec 24Feb 25Apr 25Jun 25Aug 25Oct 25Dec 25Feb 26Apr 26Jun 26Sep 2642.1K
Low 17High 57.1KAverage 20.8K
Hover, tap, or use ← → to inspect a month

Backlinks. Source: DataForSEO

Ready to make a call on FantasyGF?

Visit the site, or compare it with the rest of the category first.

In short

What I learned, and what I expect next

Pulling it together, from my own paid test, from what users say, and from my indicators: FantasyGF is growing right now. Here is the whole picture in a few lines, then the scorecard behind it.

How it is performing

FantasyGF drew an estimated 877 visits from Google search in September 2026, +13.5% on August 2026 and +87600.0% on a year earlier. The year-on-year figure flatters it because it is measured against the September 2025 low; against the July 2025 peak it is at 43850%. Over the last quarter the trend is growing (+92.8% against the three months before). Counting all traffic, it is #13 among the AI girlfriend sites I track, behind GirlFriendly.ai and GirlfriendGPT.

Reading the trend with the reviews in hand: the traffic line is rising and the reviews explain both why it rises and why it does not rise faster. The feature set, images, short video, voice messages and calls at a $5.83-a-month annual price, is what gets the brand onto the list posts that bring its visitors. The execution, voice that failed for the one tester, calls at 8 tokens a minute, chat that over-writes, is why nobody on those lists comes back to describe using it and why brand searches sit flat at 1,300 a month while the rival names next to it in Trends are breaking out. The five-week Tranco jump is the number to re-check in November; growth bought with ads shows up there first and fades there first.

What changed

Brand demand is the standout: about 1,300 US searches a month (+30.0% year on year) and 4,400 worldwide. Demand and traffic are rising together, which is the healthiest pattern there is. In popularity terms it has climbed 376,433 places since August 31, 2026. Links are being lost (-15.3% in 12 months). On the reputation side: of the 3 most-watched videos whose transcripts I read, 1 is positive, 1 is a neutral walkthrough with no verdict and 1 is negative. The one hands-on video found voice and calls broken and the chat flat, while the two promotional videos list features without testing them. Reddit treats FantasyGF as a name on lists, rated level with Candy AI and described as a virtual dating app, with no user in any thread describing what it is like to use.

What I expect

With demand, traffic and links moving the same way, I expect FantasyGF to keep gaining ground over the next few months. The thing that would change my mind is a month-on-month drop that repeats, which is why I refresh these figures monthly. 8 of its 8 closest search rivals currently get more US search traffic than it does, which is the gap to watch.

Scorecard

Each indicator, its latest reading, and which way it is pointing.

▲
Search traffic, latest monthvs August 2026
+13.5%
▲
Search traffic, last quarterlast 3 months vs the 3 before
+92.8%
▲
Search traffic, year on yearvs September 2025, a low point, so this flatters it
+87600.0%
▲
Brand demand (US searches)vs September 2025
+30.0%
▼
Google Trends interest55% of the 12-month peak
55 / 100
▲
Popularity rank, 30 days#726,988 worldwide
up 376,433 places
▼
Referring domains, 12 months1,283 sites link to it
-15.3%
●
Public attention485 combined video views
0 articles · 3 videos
●
YouTube reviews, by transcript3 transcripts read, 1 neutral
1 positive · 0 mixed · 1 negative
●
Reddit threads, by stance3 threads read with top comments
1 positive · 0 mixed · 0 negative

▲ improving · ▼ worsening · ● steady or not directional. Thresholds: a change of more than 3% counts as a move.

What I will check next month

  • Whether September 2026’s traffic figure held, rose or fell once the next month is in.
  • Whether US brand searches move out of the 1.3K band they have sat in.
  • Whether the Tranco rank keeps the direction it showed over the last 30 days.
  • Whether fantasygf.com holds position #80 for “ai girlfriend”, the category’s own search.
  • Whether the rival names rising next to it in Google Trends (herahaven, nectar ai, ai girlfriend) keep climbing while its own searches hold.

Part 7 · Method

How I measured all this

I bought a FantasyGF membership and tested it myself, and I come at the write-up as a data researcher. Everything on this page is measured, because the whole point of this site is to rank AI girlfriend platforms by real usage and interest rather than by anyone’s taste, and every brand page uses the same indicators in the same order so they can be compared. I pulled every figure on October 7, 2026.

Traffic, keywords, backlinks, competitors, WHOIS
From DataForSEO, refreshed monthly. “Organic traffic” is an estimate of visits from Google results, modelled from the keywords a site ranks for. It leaves out direct visits, app traffic, paid ads and social referrals, so the real audience is bigger; that is why I also show an all-traffic popularity rank.
Brand search volume
Google Ads keyword data via DataForSEO. Google rounds these into bands, which is why months often repeat the same figure.
Domain Rating
Domain Rating by Ahrefs, a 100-point measure of backlink strength, published under Ahrefs’ Domain Rating license. The DataForSEO domain rank is a separate 100-point backlink score.
Popularity rank
The Tranco list, an academic ranking of the top million domains built from DNS and browser data; a rank, not a visit count. The Chrome UX Report tier is Google's own popularity bucket from real Chrome users.
Trends, videos, forums, press, search suggestions
Google Trends is relative interest on a 0 to 100 scale, not a count. News, YouTube, Reddit and autocomplete data come through SerpApi from public Google and YouTube results; video views are summed across the videos I found. For the most-watched videos I also pull the caption transcript through SerpApi, read it in full, and write a short account of what the video says plus my own conclusion; videos without captions are listed by title only. For Reddit I read each thread Google surfaces directly from reddit.com in a browser, record its upvotes, comment count and top comments as they stood that day, screenshot the post as it is publicly visible, and classify it after reading; threads with no analysis yet are listed by title. Press articles are read in full and classified the same way; pages that block automated readers are marked as unread. The category search check is a live Google result for the category's own term (United States, desktop) via DataForSEO. Discord counts come from the server's public invite page.
Hands-on testing
I purchased a FantasyGF membership and used the product myself: chat, memory, images and the paid extras. My own test tells me what each feature does; the verdict on each feature comes from what users publish, because for a brand this size their view is more grounded than one reviewer's.
Rounding and independence
Figures are rounded for display; charts show the exact values. Top AI Girlfriends is not affiliated with DataForSEO, Ahrefs, SerpApi, Google, or FantasyGF.

Part 8 · Final verdict

One last word on FantasyGF

FantasyGF is a full-featured mid-market companion site that does a bit of everything and nothing memorably. In my month the chat stayed in character better than the one YouTube tester allowed, images matched the companion I built, the hugging and kissing clips were a real extra, and voice calls worked, which they did not for him. The $69.99 first-year price with 100 tokens a month is one of the cheapest ways into a site with this many features. That is why it sits on every comparison list, rated level with Candy AI on Reddit.

The trade-off is breadth against polish. Voice is ordinary text-to-speech and the one outside test caught it broken; calls cost 8 tokens a minute, so Premium's allowance buys twelve minutes; the chat leans on florid lines you have to tune out; and the first-year price triples to $25 a month if you stay monthly. Companions you build are public unless you change a setting, which the marketing's 'end-to-end privacy' line does not mention. None of it is disqualifying; all of it is why nobody has made a fan video.

Who should pick it: someone who wants images, short clips and the occasional voice message from a companion they design, and will pay $69.99 up front for a year. Who should not: anyone who wants the voice to be the main event, anyone who wants a deep memory, and anyone paying month to month, where $25 buys more elsewhere.

Where it stands today: #13 of the AI girlfriend sites I track, growing on my scorecard, with a Tranco rank that improved by nearly 400,000 places in five weeks, search traffic under a thousand visits a month, and brand searches flat at about 1,300. It climbs my ranking if the reach gain holds into November without the brand term falling; it is a mid-table site with a good price until then.

See where FantasyGF stands among every AI girlfriend site

Every platform I track, tested the same way and ranked by the same indicators.