Your newsletter may get clipped.
Gmail's Gemini AI now deprioritizes newsletters it detects as AI-written. Complete survival guide for email creators and newsletter publishers in 2026.
It started with a weird Tuesday. Marcus runs a marketing newsletter called The Weekly Shift. He has been publishing for three years. His list sits at just over 4,000 subscribers, built slowly through LinkedIn shares, a few viral posts, and relentless consistency. His open rates had been a steady 28 to 32 percent for most of 2025. Nothing extraordinary, but solid. He was proud of it.
In early February 2026, he checked his analytics after sending his usual Wednesday issue. Open rate: 19 percent. He checked again an hour later. Still 19 percent. He figured it was a slow morning. By Thursday, it had climbed to 21 percent. That felt like a slump, not a crisis. He had heard about winter subscriber fatigue. He moved on.
The following week: 18 percent. The week after that: 17 percent. Then the email arrived. A reader he had heard from before, a product manager named Sarah, sent him a short message. "Hey, just wanted to let you know your last three newsletters went straight to my spam folder. I only found them because I was cleaning it out. Not sure if you know." He did not know.
Marcus went back and looked at the timeline. The drop started on January 29. He thought about what had changed around that date. He had been doing the same things: same subject line format, same Beehiiv setup, same send time. But then he remembered. Around January 20, he had started using an AI tool to draft his newsletters. He would give it bullet points about the week's marketing news, it would write a polished first draft, and he would skim it and hit send. He thought he was working smarter.
Marcus's story is showing up in email marketing forums, newsletter Discord servers, and creator communities everywhere in 2026. Open rate drops of 30 to 50 percent. Spam folder complaints from longtime subscribers. A timeline that maps precisely to when the sender switched to AI-drafted content. This is not a coincidence or a seasonal trend.
What Marcus experienced is the result of Gmail's Gemini AI system doing exactly what it was built to do: score incoming email for authenticity, personal relevance, and genuine human communication. Newsletters that fail that scoring get buried. Not marked as spam necessarily, not bounced. Just quietly dropped to the bottom of the Promotions tab, or worse, routed to spam for enough users that your sender reputation starts to erode. And once that spiral starts, it compounds fast.
Most newsletter creators think of email filtering as a binary: spam or not spam. That mental model is about five years out of date. Gmail's current architecture does not just sort email into two buckets. It runs a continuous scoring system that affects where every email lands, how prominently it displays in the inbox, and whether Gmail's interface actively surfaces it or quietly depresses its visibility.
Gemini's integration into Gmail, which rolled out in phases through 2024 and became fully operational for free accounts in late 2025, added a new scoring layer on top of the existing spam filters. The existing filters were good at catching phishing, malware, and obvious bulk spam. They were not designed to evaluate writing quality, authenticity, or whether a newsletter sounded like a real person wrote it. Gemini is.
The technical reality is that Gemini processes email content before delivery classification. It reads the body text, the subject line, the preview snippet, and metadata about the sender. It scores the email on several dimensions, including what appears to be a measure of personal communication authenticity. That score feeds into Gmail's classification logic, which then determines whether the email goes to the Primary tab, Promotions tab, Updates tab, or spam, and within any given tab, where it appears relative to other emails.
A lot of newsletter creators shrug about the Promotions tab. "My readers know to check it," they say. This was never fully true, and in 2026, it is dangerously false. Gmail's interface has changed. The Promotions tab now has its own internal ranking system. Emails that score higher get shown first and are displayed with a visual preview card. Emails that score lower are collapsed or pushed below the fold. A newsletter that lands in Promotions but scores poorly might as well be in spam for most of your subscribers.
| Placement Tier | Typical Open Rate | What This Means |
|---|---|---|
| Primary Tab | 28-35% | Treated as personal communication — strongest signal |
| Promoted Promotions (visual card) | 15-20% | Top-ranked emails with visual preview card |
| Standard Promotions | 8-12% | Visible but below fold on mobile |
| Deprioritized Promotions | 3-6% | AI-flagged content, buried at bottom |
| Spam Folder | 4-8% | Subscriber manually finds it or never does |
Deprioritization is not the same as spam classification. When your newsletter gets deprioritized, your deliverability score stays intact on paper. The email reaches the inbox. Gmail counts it as delivered. Your email service provider shows a successful send. But Gmail's presentation of that email is suppressed. It sits lower in the tab stack. It does not get the visual preview card treatment. For mobile users, who make up roughly 60 percent of Gmail opens, the email may not appear on the first screen of the Promotions tab at all.
This is why newsletter creators often have no idea it is happening. Your technical deliverability metrics look clean. Your list is not shrinking. But your opens are dropping because Gmail is silently choosing to show your email to fewer people, or to show it in a position where they will not see it without actively scrolling.
The volume of AI-drafted newsletters exploded in 2025. As every creator with access to ChatGPT, Claude, or Gemini itself started using AI for first drafts, Gmail's inbox became flooded with content that had a specific texture: polished, structured, and completely devoid of the quirks and specificity that genuine human communication carries. Google built Gemini's email scoring partly to address this exact problem. The filter is not an accident. It was a deliberate product decision.
What Gemini appears to evaluate falls into three broad signal categories: linguistic patterns, structural patterns, and engagement metadata. No single signal triggers the filter. It is the combination and concentration of signals that determines your score.
AI-generated text has a set of linguistic fingerprints that are difficult to eliminate entirely if you are not specifically trying to address them. Overly smooth transitions between ideas. Consistent paragraph length. Absence of sentence fragments or idiosyncratic punctuation. A tendency toward complete, grammatically perfect sentences where a human writer would naturally use contractions, get informal, or trail off. Subject-verb agreement that is almost robotically consistent. The use of framing phrases that LLMs favor, like "it is important to note that" or "one key consideration is." These patterns show up reliably in AI-drafted text and Gemini was trained on enough text to recognize them.
What Gmail rewards on the linguistic side is the presence of genuinely personal communication markers. Sentence fragments that would never make it through a grammar check. Informal asides in parentheses. Opinions stated without hedging. Self-referential observations. References to specific dates, specific people, or specific places that narrow the content to one particular person's experience. A human newsletter writer might say: "I tested this last Thursday and it blew up in my face, which tells you something." An AI newsletter would say: "Testing revealed challenges in implementation, suggesting areas for improvement." Both convey a similar fact. Only one of them reads as human.
AI tools love structure. They love to produce newsletters with a clear intro section, a main body divided into numbered or bulleted sections, and a brief closing call to action. That structure is readable and useful. It is also a structural fingerprint that Gemini has clearly been trained to recognize. The "three tips and a CTA" format is probably the single most flagged newsletter template in 2026. It is everywhere, it is predictable, and Gemini knows it.
Here is the part that most people miss: engagement signals can override content signals. If your readers consistently open your emails, click your links, and reply to your messages, Gmail's system interprets that engagement as validation that your content is wanted and relevant, regardless of how it was written. Strong engagement is essentially a vote from real humans that your newsletter is worth their time. Gemini respects that vote.
This is both good news and a complication. The good news is that if you have a genuinely engaged list, you have some protection. The complication is that if your AI-drafted newsletters have been landing in Promotions for a few months and your engagement has been drifting down, you are in a compounding negative cycle. Lower engagement leads to lower scoring, which leads to worse placement, which leads to even lower engagement.
Your domain and sender reputation act as a multiplier on your content score. A sender with a strong reputation gets the benefit of the doubt on borderline content. A sender whose reputation has been eroding, due to declining opens, spam complaints, or poor list hygiene, gets less benefit of the doubt and needs cleaner content signals to compensate. Do not neglect your sender reputation while focusing only on content fixes.
The roundup format, where you gather five to ten links from the week and write a brief paragraph about each one, is probably the most common newsletter format among creators who switched to AI drafting. It makes sense: give the AI a list of links, have it write a summary sentence or two per item, add a subject line, done. The problem is that this format produces almost pure AI-linguistic fingerprints. The summaries are grammatically uniform. The transitions between items are identical in structure. There is no consistent voice across the items because the AI is synthesizing from each source individually rather than filtering everything through a single perspective.
There is a fundamental difference between a newsletter and a blog post. A blog post is written to be consumed by a stranger who found it through search. A newsletter is written to a specific audience that signed up for your perspective. AI tools, by default, write long-form content that sounds like blog posts, because that is mostly what they were trained on. The result is newsletters that have H2 headers, numbered sections, and a "conclusion" paragraph, all of which are structural signals that say "blog post" rather than "personal communication." Gmail's Gemini system is not fooled by the fact that it arrived via email.
AI-generated subject lines tend to be descriptive and complete. "Five Marketing Strategies That Will Transform Your Q2 Results." "The Complete Guide to Understanding Your Email Analytics." These subject lines pattern-match to broadcast commercial email, not to personal messages. Real newsletter subject lines are often incomplete, curious, or oddly specific. "This surprised me." "The number was 3,200." "I was wrong about this." The more your subject line reads like a headline, the more it reads as AI-generated broadcast content.
If your newsletter looks like: intro paragraph, tip one, tip two, tip three, CTA to buy or read or click something, that format is getting flagged hard in 2026. It is not that the format is inherently bad. It is that every AI assistant on the planet defaults to producing it, and Gmail has seen it millions of times. The format has become so associated with AI-generated promotional content that even human-written newsletters in that format may take a scoring hit simply by association.
If your newsletter targets professionals who primarily read it on Gmail during work hours, your AI content penalty is likely higher than average. Your audience expects a professional but personal voice, and they are reading in a context where Gmail is especially sensitive to the distinction between genuine business communication and broadcast promotional content.
You do not have to stop using AI tools. That is not the conclusion here. The conclusion is that you cannot use AI tools the way most people have been using them: ask for a draft, skim it, send it. That workflow produces content that reads as AI, and Gmail's Gemini system now penalizes that content. The fix is not less AI. The fix is better editing and better writing practices that inject the signals Gemini rewards.
Voice injection is the process of taking an AI-drafted newsletter and adding the specific signals that make it read as personal human communication. It is not about changing the content. It is about changing the texture. The technique works in three passes.
First pass: add your opening. Do not use the AI's opener. Write your own two to three sentence opening that references something specific that happened to you this week. Specific date, specific situation, specific observation. "Tuesday afternoon I got a Loom from a subscriber asking why my last issue went to her spam folder. I had no good answer. Turns out I had been making the exact mistakes this issue is about." That opening is 35 words. It contains a specific day, a specific communication format, a specific subscriber interaction, and a specific personal admission. Gemini reads that as human.
Second pass: add your opinion. Go through the AI draft and find every place where it presents information neutrally. Replace neutral presentation with your actual take. "Some marketers prefer shorter subject lines" becomes "Short subject lines almost always outperform long ones, and if someone tells you otherwise they are looking at the wrong metric."
Third pass: add specificity. Find every place where the AI used a general example or a hypothetical and replace it with something real and specific. "For example, a retail company might..." becomes "For example, a Shopify brand I spoke with last week..." The specific context makes the content feel like it came from someone who exists in the world and has conversations with actual people.
The single most effective structural change you can make to a newsletter is to write it like a letter to a specific person rather than an article for a general audience. This means writing "you" constantly. Addressing the reader directly. Assuming shared context and referring to it. Writing like the reader already knows you, because they signed up for your list, which means they do.
In practice, this means dropping the header hierarchy. No H2s, no H3s, no numbered sections in the email body. Just paragraphs, with natural transitions. No bullet points unless they are truly lists, not just structured information that the AI converted to bullets for readability. Start sentences with "And" and "But." Use contractions. Write "I am" as "I'm." Include asides in parentheses. These small style choices collectively produce a very different linguistic signature from AI-generated content.
Read your subject line and ask: could this be the subject line of a personal email from a friend? If yes, it scores well. If it reads like the headline of an article or the subject of a marketing email, rewrite it. The subject line should tease a conversation, not promise a piece of content.
Content signals matter, but engagement signals can override them. If enough of your readers are opening your emails, clicking your links, and replying to your messages, Gmail interprets that as definitive evidence that your newsletter is wanted.
Before your next newsletter send, segment your list by engagement recency. Pull out the subscribers who have opened or clicked in the last 90 days. That is your re-engagement segment, your most valuable audience right now. Send your next two or three issues exclusively to that segment. This does two things: it improves your aggregate open rate immediately (because you are only sending to people who have recently been active), and it sends a strong signal to Gmail that your email gets opened when it reaches engaged readers.
Every newsletter should end with a specific, easy reply prompt. Not "let me know what you think!" That is too vague and too familiar as a boilerplate CTA. Specific questions that take 30 seconds to answer. "If you could change one thing about how you are currently writing newsletters, what would it be? Reply and tell me." or "Which section of today's issue was most useful? Just say '1' or '2' or '3' in a reply." Low friction, high specificity.
Beyond the direct engagement value, replies do something else: they train Gmail to put your newsletter in Primary tab for the people who reply to you. When a subscriber replies to your newsletter, Gmail's system notes that this sender and recipient have a two-way communication relationship. Future emails from you to that subscriber are significantly more likely to land in Primary rather than Promotions.
| Platform | IP Reputation Pool | Default Email Format | AI Detection Risk | Deliverability Notes |
|---|---|---|---|---|
| Beehiiv | Shared with verified creators | HTML with clean structure | Medium | Good deliverability, but shared IPs mean reputation spillover from other senders |
| Substack | Substack-managed | Minimal HTML, near plain-text feel | Lower | Substack emails often land in Primary due to perceived personal communication format |
| Mailchimp | Large shared pool | Heavy HTML templates | Higher | Heavy HTML is a promotional email signal; requires extra effort to score as personal |
| ConvertKit | Shared with vetted senders | Plain text default available | Lower to Medium | Plain text option significantly reduces AI detection risk |
Substack has an inherent deliverability advantage in the Gmail era because its default email format, minimal HTML that resembles a personal email, matches exactly what Gemini rewards. If you are on Substack and struggling with Gmail deliverability, the platform is probably not your problem. If you are on Mailchimp with a heavy HTML template full of images, headers, and branded design elements, you are working against yourself on the Gmail scoring dimensions before you even write a word.
Plain-text newsletters, or very minimal HTML that resembles plain text, score significantly better on Gmail's personal communication dimension than rich HTML newsletters. Deliverability data from 2025 and 2026 consistently shows plain-text or near-plain-text newsletters achieving Primary tab placement at rates two to three times higher than equivalent HTML newsletters from the same sender. The visual richness that HTML enables comes at a real deliverability cost in the Gemini era.
The most common mistake is using an AI tool to write the newsletter from scratch and then light-editing for accuracy without changing the voice or structure. A quick proofread does not change the linguistic fingerprint. The fix is the voice injection technique: three full editing passes that add specific personal details, explicit opinions, and a letter-format structure.
Most newsletter creators check their open rates after each send and move on without looking at the trend. A single low open rate looks like a bad week. Four consecutive declining open rates is a scoring problem that is getting worse. If your Gmail opens are declining while your non-Gmail opens are stable, that specific pattern is the Gemini filter signature.
Templates like "This Week in [Your Topic]: [X Things You Need to Know]" are open rate killers in 2026. They are recognizable as newsletter broadcast formats. Every time you use a template subject line, you are competing against every other newsletter using the same template, and you are signaling to Gemini that this is promotional broadcast content rather than personal communication.
Newsletter creators who treat their newsletter as a broadcast medium, sending content to readers without ever expecting or requesting a response, are missing the biggest engagement signal available to them. Replies are the strongest single indicator Gmail has that a communication is genuine and two-directional. Build in a specific, frictionless reply prompt in every issue.
Sending your full list every week, including subscribers who have not opened in a year, is a self-inflicted deliverability wound. Every non-open from an inactive subscriber drags your engagement rate down, which feeds back into your Gmail scoring. Suppress subscribers who have not opened in 90 days for your main sends and run a separate reactivation campaign for them.
Paying a designer to create a beautiful HTML newsletter template with your logo, branded colors, a header image, and a structured three-column layout feels like professionalism. In the Gmail Gemini era, it is a deliverability penalty. Heavy HTML email templates are one of the strongest signals that an email is promotional broadcast content rather than personal communication.
Preview text is the 40 to 80 characters that appear next to your subject line in the inbox view. Most creators do not set it, which means their email platform auto-generates it from the first line of the email body. In an AI-drafted newsletter, that first line is often a topic introduction sentence that reads exactly like the opening of a blog post. Setting your preview text to something personal and conversational costs you nothing and improves your scoring.
Audit your last 30 days of Gmail deliverability
Before writing a single word, understand your current situation. Log in to Google Postmaster Tools and check your domain reputation score. Segment your last four to six newsletter sends by email provider and compare your Gmail open rate to your non-Gmail open rate. If the gap is larger than 5 percentage points and declining, you have a Gemini filter problem that needs addressing before your next send.
Clean your list before the next send
Export your subscriber list and filter for last-open date. Suppress anyone who has not opened in 90 days from your main send list. Create a separate segment for 90 to 180 day inactive subscribers for a reactivation campaign. Remove anyone who has not opened in over 180 days entirely. A smaller, more engaged list scores better with Gmail.
Write your personal opening before touching AI
Before you open an AI tool, write two to four sentences in your own voice about something specific that happened this week that connects to your newsletter topic. Specific date, specific situation, specific reaction. This does not need to be polished. It needs to be real and specific. Do not delegate this section to AI.
Use AI to draft the body content only
Now open your AI tool. Give it a clear brief: the topic, the specific points you want to cover, and any source material. Ask for a draft of the main body content. Do not ask it to write the opening, the subject line, the closing, or the reply prompt. Let it draft the substantive middle section.
Run the three-pass voice injection edit
Pass one: read every paragraph and add or replace with your specific opinion on the topic. No neutral statements allowed. Pass two: find every general example and replace with something specific and real from your actual experience. Pass three: check paragraph length variation. If every paragraph is 3-4 sentences, break some up. Add a one-sentence paragraph somewhere. Make the rhythm irregular.
Write a personal-format subject line
Write five subject line options. None of them should describe the contents of the newsletter. All of them should feel like the beginning of a sentence from a friend. Under 50 characters when possible. No Title Case. No numbered lists in the subject line itself. Pick the one that creates the most curiosity while still being honest about what is inside.
Set your preview text manually
Do not let your email platform auto-generate preview text from the first line of your email. Set it manually to something that extends the subject line conversation. Keep it under 70 characters. Make it conversational. Make it feel like the second half of a spoken sentence.
Check your From name and reply-to address
Confirm that your From name is your personal name, not your newsletter brand name. Confirm that your reply-to address is a real monitored inbox, not a no-reply address. No-reply addresses are a strong broadcast-email signal and they prevent the reply engagement that is one of your most valuable deliverability assets.
Add a specific reply prompt near the end of the body
Before your closing CTA, add a specific question that takes 30 seconds to answer and feels genuinely curious, not performative. End with a direct prompt: "Hit reply and let me know." This is not optional decoration. Every reply you receive is a positive signal to Gmail for every future email you send to that person.
Run a final AI-detection check before sending
Before you hit send, paste your newsletter body into an AI detection tool and check the score. If it comes back at over 50 percent AI-generated, do another editing pass before sending. Most deliverability experts recommend getting below 30 percent on mainstream detectors as a rough benchmark.
Send to your engaged segment first
For the next four to six sends, send to your engaged segment (90-day active openers) first. Monitor your Gmail opens specifically. Once your Gmail open rates start trending back up, you can gradually reintroduce your reactivation campaign subscribers.
Monitor and iterate weekly
After each send, check your Gmail-specific open rate, your click-through rate, and your reply count. Track them in a spreadsheet. Look for trends over four to eight weeks. The improvements from better content and list hygiene are not instant. They accrue over multiple sends as Gmail's scoring system updates your sender reputation.
Filtered subject line: "Five Email Marketing Tips to Boost Your Open Rates in 2026." Body opens with: "Email marketing remains one of the most effective channels for reaching your target audience. In this edition, we will explore five proven strategies to improve your open rates and drive better results." Three tips follow with H3 headers and bullet points. Closes with: "Implementing these strategies can significantly improve your email marketing performance. Feel free to reply with any questions." Gmail classification: Promotions tab, lower visibility tier.
Promoted subject line: "the email subject line mistake I keep seeing." Body opens with: "I reviewed 12 newsletters from my subscribers last week, and 9 of them were making the same subject line mistake. I have made it too, so I am not throwing stones. But it is costing people real opens and I want to show you exactly what I mean." The same five tips follow, but written as personal observations with specific examples the writer encountered. Closes with: "Which of these felt most relevant to where you are right now? Reply and tell me, I read every response." Classification: Primary tab for subscribers who have previously engaged, Promotions top tier for others.
Filtered subject line: "This Week in SaaS: Funding Rounds, Product Launches, and Market Trends." Body is structured with three to five sections, each covering a different news item. Each section is two to three paragraphs summarizing the news with a neutral, informative tone. No personal perspective. No specific opinion on what the news means. Classification: Promotions tab, often low visibility.
Promoted subject line: "the funding round that surprised me most this week." Body opens with: "Three interesting things happened in SaaS this week. One of them I predicted. One of them I completely missed. And one of them I think everyone is misreading." The same news items follow, but framed through the writer's specific reactions. Closes with: "Disagree with my read on the third one? I suspect some of you will. Tell me why." Classification: Primary tab for active readers.
Filtered subject line: "How to Build a Successful Digital Product: A Step-by-Step Guide." Body is a long-form how-to article with multiple H2 headers, numbered steps, and callout boxes. Total word count around 1,500. Heavy HTML template with logo header and footer links. Classification: Promotions tab bottom tier, occasional spam classification.
Promoted subject line: "I launched 4 products. 3 failed fast. Here is what the 4th one did differently." Body is written as a personal narrative in plain text. No headers. Just paragraphs. The same information about product building is present, but delivered as a story about what the writer actually experienced: what they tried, where they failed, what specific thing they changed on the fourth product. Closes with: "What stage are you at right now with your product? I am putting together a resource specifically for wherever people are stuck. Reply with your situation." Classification: Primary tab for engaged subscribers.
In every case, the promoted version contains the same core information as the filtered version. The difference is entirely in delivery: personal opening, specific details that could only come from one person, explicit opinions, and a reply invitation. None of this is about being more creative or more entertaining. It is about writing like a person communicating with a specific audience, not like a system broadcasting content to a list.
Open Google Postmaster Tools and check your domain reputation. If you are not already set up there, set it up now. It takes ten minutes and it gives you the ground-truth data on how Gmail currently views your sending domain. If your reputation is "High," you have a content problem that is fixable with the voice injection technique. If your reputation is "Medium" or lower, you have both a content problem and a reputation recovery problem, and you need to address both simultaneously.
Then look at your last newsletter. Read it out loud. Notice where it sounds like an article and where it sounds like a person. Notice where you state an opinion and where you present neutral information. Notice whether there is a single specific detail in the entire email that could only have come from you and your specific experience last week. If you read it and it could have been written by anyone with access to an AI tool and knowledge of your topic, that is exactly what Gmail thinks too.
Marcus rewrote his workflow. He still uses AI for research and first drafts. But he writes his own opening every week, references something real and specific that happened, states his actual opinions on the news he covers, and ends every issue with a question his readers actually answer. His Gmail open rates are back above 25 percent. His reply rate tripled.
Gmail's Gemini system is filtering for personal communication, not gatekeeping against AI tools. You can still use AI. You cannot send raw AI output. Write your own opening with a specific day and a specific situation. Add your opinion where the AI was neutral. Replace generic examples with real ones from your actual experience. Ask for replies with a specific, low-friction question. Clean your list. Move to simpler HTML. Do this consistently for four to eight weeks and your Gmail deliverability recovers. Keep sending generic AI drafts and it keeps getting worse. The filter is working as intended.
HumanLike helps you turn AI-drafted newsletters into content that reads as genuinely personal, the kind Gmail's Gemini filter promotes, not buries.
Writing about AI humanization, detection accuracy, content strategy, and the future of human-AI collaboration at HumanLike.
On This Page
Discuss this article with AI
Riley QuinnTurnitin's August 2025 update silently killed every bypass method that was working. Detection rates spiked overnight. Here is the full breakdown of what changed technically, which strategies are now dead, and the exact methods that still pass in 2026.
Riley QuinnGPT-5 is a better writer than GPT-4. It is also harder to disguise. The same qualities that make it impressive, ultra-consistent prose, near-perfect structure, flawless grammar, are exactly what modern detectors are trained to spot. This guide breaks down why GPT-5 triggers detection systems harder than its predecessors and gives you the full workflow to fix it.
Steve VanceClaude Opus 4.6 produces some of the most sophisticated AI-written text available in 2026. It also has one of the most recognizable detection signatures. Long hedging chains, philosophical asides, stacked qualifications, and words like 'intricate' appearing in predictable positions make Opus output almost trivially identifiable to modern detectors. This guide covers everything: what makes Opus detectable, how its signature differs from GPT-4o, what Turnitin and GPTZero specifically flag, and the complete workflow to humanize Claude Opus output using humanlike.pro.