What shipped, newest first.

The Instagram and TikTok comment viewers load the most liked comments. On a big post they now show it above the list, for example Top 100 of 323,062 comments, and a search with no match says it only looked at those loaded comments.

The YouTube and TikTok comment scraper pages now open with the free viewer. Paste a video link and read the comments right there, with search and CSV export, before you sign up for the ranked ad angles.

get_video_transcript now reads the audio itself when a TikTok has captions turned off or an Instagram video runs longer than 2 minutes. Before, those videos came back with no transcript. It still costs 1 credit per transcript, and videos with no spoken words stay free.

Each platform now has its own page with the questions that work in Claude, what it costs and how to connect: Reddit threads, YouTube comments and transcripts, TikTok comments, Trustpilot reviews and LinkedIn post comments.

New accounts paste a link or describe the product, and Adlicio finds the threads, reads the comments and shows the first ad angles with real quotes in about a minute. The full report keeps building in the background, and a 7-day free trial of the monthly plan sits right next to the angles.

On a phone, the setup prompt now shows in full with a big copy button, quotes fly into the speech bubble one at a time, and tapping the niche box no longer zooms the page. The tool keys stay in place between demo runs, and the chat button is a small avatar that waits until you scroll before it says hello.

The top bar of the site now has the same six tools as the app: Scrape, Brands, Radar, Hooks, Scripts and History, each with its animated icon. Platforms lists every site Adlicio scrapes. On a phone, the menu shows the tools as a grid.

The home page opens with a live demo: your agent asks for ad angles, Adlicio pulls real comments from Reddit, Amazon, TikTok and more into a 3D speech bubble, and the top angles rank on screen. Try your own niche in the agent box, or copy the one-line setup prompt for Claude, ChatGPT or any agent. It follows your light or dark setting.

A new All tools row in the sidebar opens every tool on one panel: Scrape, Brands, Radar, Hooks, Scripts and History, each with its pages, plus the Chrome extension, Claude and ChatGPT, the CLI and the API. Each tool has its own animated icon that plays when you point at it.

Paste a Google Maps Share link (maps.app.goo.gl) into Claude or ChatGPT and Adlicio opens the place and reads its reviews; before, only the full /maps/place/ link worked. The docs add setup pages for Codex and VS Code, every input of every tool with its type, default and limits, a plans table, and a troubleshooting entry for each error message.

The developer docs, the MCP docs and the guides are now one site at /docs, with search (⌘K), tabs for Documentation, Tools & sources and CLI & API, and a separate page for every topic. Each MCP tool has its own page with its inputs, price and plan. Copy any page as Markdown or open it in Claude or ChatGPT. Old links still work.

The help center has 8 guides, one job each: add Adlicio to Claude or ChatGPT, research Reddit, see a competitor's Facebook ads, transcribe a TikTok, Instagram or YouTube video, analyze Amazon reviews, pull TikTok comments and read Google Maps reviews. Each one has a prompt to copy, a real result and the current price.

On a phone the app has a bottom bar for Home, Find, Brands and History, and More opens every other page, your work, credits and help. Brand tabs show three and the rest under More, the History mark bar and Radar's reply sit at your thumb, and the History reader floats over the list on tablets. Small motion now confirms what happened: results rise in, a new mark sweeps, the work count rolls. It stays still if your device asks for less motion.

Brands shows how old each brand's research is, with Refresh research and its price on the brand that needs it. Hooks puts the hook generator, Scroll stoppers and Creator finder in tabs. Scripts is a list beside a reader, with filters kept in the link. Boards opens on templates, Reports on one list with a chip per type and a New report menu with prices, and Saved videos says plainly how to fill it.

Radar shows its posts as a list beside a reader with your reply, Draft reply and the keyboard keys in view; Skip and Replied can be undone. Listening starts with one field to watch a link, subreddit or keyword, suggests places from your research with the segment that talks there, and keeps the weekly report beside it. Connected accounts lists all 9 platforms with Connect on each.

A brand's Overview now shows who the buyer is, the segments, a lead quote and the angles, with Write script on each, plus your marked lines, where the sources come from and what Listening watches. Research reads as one brief: pick a run and a segment, open an angle, edit its hook and test, then create an ad or write a script. The 8 tabs follow the work, and every button that spends credits shows its price.

History opens a thread in a reader beside the list, so you keep your place. Select a line and press H to mark it: the line keeps its link and lands in the brand you pick, and the new Marked tab lists every marked line. Rows are grouped by day, and Type, Source, Brand and Made by filter the list.

Scrapes, brand research, deep scrapes, browser pulls, Instagram follower pulls and reports now share one tray in the sidebar: what is running, what needs you, and what is ready, with one next step each. Home lists what landed since your last visit, including work from Claude and new Radar posts to triage.

Type a topic in ⌘K to find discussions, watch it with Radar or write hooks, each with its price. Paste links to scrape them. Saved work is found by title, every page is still in Go to, and ⌘Enter runs an action in the background with a note when it lands.

Type a topic or paste links on Home and press Enter: the search runs and opens Find and scrape with the best 6 picked. Every search is kept under Recent runs, including searches from Claude and ChatGPT, and reopens with its results. Scrapes keep going on our side, so you can leave the page, stop them, or retry one link.

App text is now at least 12 px. Buttons that spend credits show the price, the brand header shows what Find sources and Run research cost, and each screen keeps one orange button for the next step. Deleting asks in the app instead of a browser pop-up.

The same 11 pages now sit under Research, Create and Saved, with counts beside Brands, History and Reports and new Radar posts. The footer folds into a few short lines, chat opens from the sidebar instead of floating over pages, and the sidebar folds to icons on narrow screens.

New accounts paste their store link first, see what Adlicio read from it and pick a first search that runs straight away. Every step has Back, the where-did-you-hear question comes last, and a new account that leaves part way returns to the same step.

When a run needs All Access or more credits than you have, its button now says so up front and offers Get All Access or Add credits, instead of failing after the click. Locked sources look locked, and Radar and Connected accounts show a lock in the sidebar on plans without them.

Video transcripts and TikTok comment pulls from Claude, ChatGPT and the app are now saved under the video's own title, not a platform and an ID. Older YouTube and TikTok rows in History got their titles too.

Your balance reads the same everywhere: one number of credits in the rail, every page header and Billing, and each header chip opens the same credits menu. Billing splits it into plan and pack credits and lists the price of every tool, taken from the numbers the server charges.

Create, edit, pause and check a Radar from Claude or ChatGPT (create_radar, update_radar, check_radar) and mark posts replied or skipped, or run commentscraper radar create, check, mark and draft in the terminal. search_reddit and commentscraper search --new also find the newest posts on demand.

Describe who you help and pick up to 10 subreddits. Radar reads their newest posts every 30 to 60 minutes, keeps the ones from people asking for exactly that, and drafts a helpful reply you post yourself. Strong fresh posts can alert you by email or Slack.

Type tryadlicio.com/ before a Reddit, Instagram, TikTok, YouTube, X, Threads or LinkedIn link and the matching free viewer opens and loads it right away.

The free Instagram comment viewer now has pages in seven languages besides English, each linked to the others, so visitors can read, search and export public comments in their own language.

A Brazilian Portuguese version of the Instagram comment viewer lets visitors read, search and export public comments without an Instagram account.

get_ad_transcript and Ad Library links in Sources now transcribe long video ads and ads without captions from their audio. A 7-minute ad works. It still costs 1 credit, and ads with no speech stay free.

search_youtube now takes a channel URL or @handle and lists that channel's latest uploads with views, upload age and length, plus the channel's subscriber count. A URL ending in /shorts lists its Shorts. It costs 0 credits on All Access.

A new page shows how to read any business's Google Maps reviews inside Claude: the setup, a real run on 212 reviews, questions that work for local businesses and what a pull costs.

A full Google Maps pull (up to about 200 reviews) now costs 4 credits. In Claude, ask for 50 reviews or fewer (max_comments) and it costs 2, the old price. The price is set before the scrape runs.

Free and Pro accounts can scrape Instagram posts in the app: 5 pulls a week, up to 100 comments each, no credits. The allowance is shared with Claude. Other sources still need All Access, and the message now names the source.

Drop your Instagram data export and get your follower list with follow dates. Search it, sort it and download a CSV. Free, and the file is read in your browser, not uploaded.

The free Influencer Finder now filters by creator size: nano (1k-10k), micro (10k-100k) or large (100k+). TikTok creators show up again once a new niche finishes collecting, in the free tool and in find_creators.

Type up to 5 keywords for your niche and Adlicio scans the Meta Ad Library a few pages deep, hides unrelated advertisers and ranks the brands by how many of their live ads are about your niche. 3 credits per scan, also in Claude as find_competitor_brands.

Click Watch on any brand in Competitor ads. Adlicio checks its live Meta ads every week and emails you when the count jumps or new ads take off. Checks are free; each alert costs 1 credit.

Competitor ad searches in the app and through the Adlicio connector now group ads by the brand behind them, show how many live ads each brand runs in your country and list its longest-running ads. Small pages with few likes often belong to brands running hundreds of ads. A new Stores only filter drops clinics and local services. A dashboard search now costs 1 credit, the same as in Claude, and a search with no ads is free.

Free accounts can now pull the comments on any public Instagram post or reel through the Adlicio connector, 5 pulls a week with up to 100 comments each, at no credit cost. Ask Claude what the audience asks, doubts or complains about.

An invite to a teammate's workspace now shows at the top of your Account page on every tab, with Accept right there. Owners see that an invited teammate has no access until they accept.

A Spanish version of the Instagram comment viewer lets visitors read, search and export public comments without an Instagram account.

Compare Instagram posts by format and TikTok videos by account size with two free calculators and sourced 2026 benchmarks.

Draw one to ten winners from up to 300 top-level comments on any public Facebook post, reel or video.

Search public Reddit comments by keyword across Reddit, inside one subreddit or thread, or through a user's comment history, then refine the matches and export a CSV.

Find discussions now searches TikTok videos, Instagram posts and X posts with real comment threads, and paste any public X post to collect its replies. A relevance filter drops off-topic hits from the noisier sources before they reach your results.

Topic search now reaches Hacker News, Blind, Trustpilot, G2, the App Store and Google Play alongside Reddit, YouTube and Quora. YouTube results come from YouTube's own search instead of a web index, so more videos match. Deep search fans the same lanes wider.

A practical guide to TypeSafe's Jev decision model with Adlicio. Three recorded workflows label real customer comments to find buying objections, recommendation requests and ad claims the evidence does not support. The source data and a Node.js recipe are free to download.

Every brand report now has a Strategy tab with ranked opportunities, a channel plan and the evidence still missing. Angles that make health or drug-comparison claims carry a claims flag with a safer framing. Quote links survive typographic apostrophes, so far more customer quotes open their original comment. Free plans always get their first report.

Paste a Meta ad, preview it in a simulated feed and get a Winner Score fitted on real Ad Library ads that kept running. Sixty yes/no questions, one second, no signup.

The comment viewer page now covers every platform: Instagram, TikTok, YouTube, X, Reddit, Threads and LinkedIn, with a section per platform and links to the matching giveaway pickers and the comment scraper. Every viewer and picker links back to it.

search_hackernews finds Hacker News stories or the individual comments that match a keyword, with points, comment counts and item links. Use it when the buyer is a developer or founder, then run scrape_url on an item link for the whole thread. Free on All Access.

get_video_transcript returns the spoken words of one public TikTok video, Instagram reel or YouTube video or Short from its link, with timed cues where the platform gives them. Read the hook, the claims and the offer a creator or competitor uses, then run scrape_url on the same link for the comments under it. A delivered transcript costs 1 credit. A video with no speech costs nothing.

Search results now say what they are: discussions, with a step line that shows find, collect, find patterns. A run that saved nothing says so and explains why, instead of reading as finished. Collecting from search opens the comments straight away. The credit meter updates the moment a search or scrape is charged. Your channels is now Connected accounts, and the locked page links to public discussion search.

Paid accounts now get a planner at the top of the handbook. Say what you sell, where customers argue about it and what you need this week, and it picks five to eight prompts in the order to run them, with your product filled in. Copy the whole chain in one click and Claude runs it step by step.

Type an X username or paste a profile or post link and read it here without an X account. A profile brings its bio, follower count and up to 60 recent posts with likes, replies, reposts and views; a post brings its newest replies, which you can search by word or username and download as a CSV.

Paste a public LinkedIn post, profile or company link and read it without logging in. A post brings its text, image, like count and public comments with a link to each commenter; a profile brings the headline, about text and recent posts; a company page brings its description, size and follower count.

Paste a TikTok video or an Instagram reel link and get the spoken words as text. TikTok transcripts keep their timestamps and download as SRT; both copy to the clipboard or download as a text file. A transcript is kept for a week, so the same link comes back at once.

Paste a public Instagram post or reel link. See about 1,000 accounts from Instagram’s public liker list. Search, filter and download a CSV. Private accounts appear by username only, without a photo.

Type a Threads username or paste a profile or post link and read it here, without a Threads or Instagram account, the app prompt or the login wall. A profile brings its bio, follower count and latest posts with likes, replies and reposts; a post brings its first replies, which you can search by word or username and download as a CSV.

Type a subreddit or paste a post, share or profile link and read it here, without a Reddit account, ads or the app prompt. Posts come with their score, flair and comment count, a thread comes with up to 100 comments you can search by word or username and download as a CSV, and adult communities are left out.

Both tools now run on the same white reading page as the Instagram viewer, with a daily budget that resets at midnight UTC instead of a weekly one, and they accept every link shape people paste, including the short share links the TikTok app copies. Each page now says exactly why a link was refused and answers the questions people ask about anonymous viewing, exports and limits.

The viewer now accepts the short share links the Instagram app copies, says exactly why a link was refused, and answers fourteen questions about anonymous viewing, exports and limits. Its shared budget is now daily, so a busy weekend no longer pauses the tool until Monday.

The free Instagram comment viewer is now a white reading page: one labelled link field, one View comments action, and a sample of what you get before you paste anything. Results arrive as a comment feed instead of a wide table, with search, sort and a CSV export that includes exactly the filtered rows, plus clear messages when a link is rejected, still collecting or empty.

scrape_facebook_group no longer requires a keyword; leave it empty to pull the newest posts from a group you can see. When Facebook's search page gives the extension no results for a keyword, the next status check queues a feed pull automatically and get_extension_job points to search_history to narrow it to your keyword.

Your channels connects your Instagram, Facebook Page, Meta ad account, TikTok, YouTube, LinkedIn, X, Reddit or Threads and lists your posts and ads with their comment counts. Pull any thread into History for free, including comments on boosted and dark posts. Adlicio only reads: it never posts, likes or replies, and DMs stay untouched. All Access only.

An Amazon pull served by the public featured-review teaser now costs 1 credit, because that path runs through a paid collection service. Pulls from the review archive or cache stay free, and a live logged-in pull still costs 2. Every Amazon pull now needs 2 credits available before it starts, and you are charged only for the tier that ran.

Paste a competitor's Instagram profile or their Meta Ad Library advertiser page onto a board and press the card's button: the board adds their latest posts or active ads as cards, one credit per expansion, each ready to scrape. A single Ad Library ad now pulls its spoken transcript from any Find and scrape surface for one credit, free when the ad has no transcript. Tall reel and TikTok cards no longer hide the cards under them inside a group.

The Instagram, TikTok and YouTube comment viewers and the giveaway pickers keep a copy of each link's comments and label it with its age. A Refresh link next to that label now pulls the newest comments right away; it counts as one run. Instagram and TikTok copies also expire after 6 hours instead of 3 days, so a post you are tracking stays current.

Two new All Access tools for ad research. get_ad_transcript returns the spoken script of a public Meta Ad Library video ad. inspect_landing_page reads the page an ad sends people to, follows its redirects, and returns the copy, headings, phone links, buttons, form fields and disclosures it finds. find_tv_ads now returns the public description and promotions of each iSpot catalog entry and scans more candidates before it sorts. find_competitor_ads accepts a minimum number of days running. A transcript or page read costs 1 credit only when it delivers; a missing transcript or a blocked page is free.

The reports page has a search box for title, brand or topic, filters for report type and status, and a live count of matches. Queued and generating reports now update on their own while the tab is open, so a finished report appears without a reload.

A link into the app now keeps its full address through the sign-in step, including the query, after a redirect, a wrong password or a switch to sign-up. The search on the 404 page opens the real thread search with your words in it.

The home page has a dedicated research input with credit costs and shortcuts to brands, scripts, hooks and boards. Mobile navigation scrolls, dialogs keep keyboard focus and Escape returns it, and the command palette finds billing, integrations, creators and scroll stoppers by description. Saved videos, listening sources and URL scrapes show a clear error when a request fails, keep your draft, and never drop a card without telling you.

TikTok comment pages now carry 50 comments instead of 20, so a default pull finishes in two calls instead of five and a deep pull reaches 500 comments. Instagram now returns the most-liked comments first, reports the post's total comment count, and includes the replies Instagram shows under each comment.

The whole price list on one page: which sources are free, what costs 1, 2, 3 or 5 credits, a worked month for one brand, and the credit packs. Rendered from the same constants the app charges, so it cannot drift.

A live Amazon pull now follows Amazon's own show-more cursor through ten pages of every filter window instead of stopping at the first page, so a large listing returns hundreds of reviews per run instead of about 140. Amazon exposes about 100 reviews per filter view, so very large listings still top out in the hundreds. The Chrome extension gets the same change in its next release.

Scraping a URL costs 1 credit on every source that runs through a paid collection service, on every plan and surface, and the monthly scrape allowance is gone. Sources Adlicio reads directly stay free: YouTube, Amazon, App Store, Google Play, Steam, Shopify stores, Hacker News, Stack Exchange and more. A live Amazon pull and Google Maps cost 2, a source search in the dashboard costs 1, and the Chrome extension stays free.

Paste a public YouTube video or Shorts URL to read, search and export its comments and replies, or choose a comment viewer from the new hub.

Pages for scheduled listening across subreddits, keywords on Reddit and YouTube, and supported review pages, with intent tags and complaint or bug alerts.

Paste a Facebook group URL, or a search inside a group, into Sources and the scrape runs through your own logged-in Chrome extension. Groups you are a member of work, private ones included, and nothing about your Facebook login leaves your browser. Turn on "Let Claude use this browser" in the extension; the job waits up to 24 hours if Chrome is closed.

Drop Reddit threads, videos, products and ads on a canvas, group them, and research every comment behind them in one place.

Four free tools that draw a random winner from the comments on an Instagram post, a TikTok video or a YouTube video: one entry per person, skip replies, require an @mention or a hashtag, exclude your own account, then draw up to ten winners in the browser and download the entries as CSV.

Click the credits meter in the sidebar to see your plan credits with the reset date, purchased pack credits that never expire, your URL scrape quota, what each action costs, and a Buy credits button that opens checkout right there, no separate page. The meter now counts plan and pack credits together, so a pack you bought shows in the number. The collapsed rail shows a compact icon with the balance.

Two new free tools: a searchable, ranked list of the 1,000 biggest subreddits with live member counts, and a review sentiment analyzer that reads up to 50 real comments from any supported review URL and returns the sentiment split plus recurring themes with verbatim quotes.

The alternatives hub now covers the Reddit listening space: Devi AI, F5Bot, Syften, KWatch, ReplyGuy, Subreddit Signals, Octolens, Brand24, Mention, Buska, Leadverse, RedReach, BillyBuzz, RedShip, Tydal, and Linkeddit, each with verified pricing and a plain answer to when the other tool is the better pick.

The free tools page now runs real searches with the receipts attached: Product Finder ranks what Reddit recommends in a category, Objection Finder lists why people do not buy, Hook Finder and Creator Finder pull from TikTok and Instagram, Competitor Ad Finder and Ad Longevity Checker read the Meta Ad Library, plus review analyzers for Amazon, Shopify, and app stores, a YouTube question finder, a Reddit mention monitor, a customer language checker for your landing page, and a free comment scraper. Every result links to its source. Free visitors get a few runs a week per tool; a free account runs them any time.

The Grok Bot window on /grok now answers one real prompt per visitor: type a question and it searches Reddit live and returns the threads it found. The second prompt asks you to create a free account.

When the shared Amazon session cannot finish a pull, the web app and the MCP hand it to your Chrome extension, which pulls the full review set with your own Amazon login. Turn on "Let Claude use this browser" in the extension.

Competitor ad cards now show the ad image or video poster, play the video inline, and can keep the mp4 in your account. The Active chip comes from the live status, not the vendor's end date. find_competitor_ads returns the same creative block.

Step-by-step pages for connecting the Adlicio MCP to Base44, Lovable, Bolt, Replit, and v0, with day-one prompts and the export API path for runtime data.

Private Facebook groups can now be researched through the user's own Chrome extension, queued from Claude with scrape_facebook_group and read back with get_history_item.

Upgrade sits in the sidebar on the free plan. Subscribers who pick another plan land on the Stripe confirm screen with the prorated amount, no detour through Account.

Watch a phrase across Reddit and YouTube on a schedule. Every new comment gets an intent tag, and complaints and bug reports raise an alert.

scrape_url reads public teamblind.com posts with replies and the poster's company. search_blind finds the posts for a keyword, ranked by relevance or recency.

scrape_url pulls the text and first comments of a public LinkedIn post. lnkd.in short links resolve on their own.

A connector page for Grok with the one-paste setup and a live thread that shows what Grok answers once it reads the real comments.

Filter search results to one platform with a chip row above the list.

A scrape can belong to several brands. The History view filters by brand on the server, and each brand page suggests threads worth scraping next.

Start from any scrape in History and get the threads that discuss the same thing, before you spend a credit.

Find the YouTube videos your audience comments on: title, channel, views and URL. Free on All Access, then scrape_url the ones that matter.

Set a language on a brand and its research, avatars and ad copy come back in that language. A German landing page lives at /de.

Plans show the numbers the app meters: research credits, monthly URL scrapes and seats. Both meters nudge at 80 percent, and the Credits Center shows the monthly scrape quota.

Give any scrape or research a custom title, inline, from the activity feed.

Find the TikTok and Instagram creators behind the videos in a niche, by country, with follower counts, engagement and public contact details. For UGC and affiliate outreach.

Overview tabs and a Credits Center that shows what is left this cycle, in one place.

One-time credit top-ups from the app, plus credit-metered overage once the monthly scrape cap is reached.

Search expands your query, runs wider lanes across sources and adds Amazon discovery, so one search covers more ground.

Select threads in History and attach them to a brand. Research briefs now quote 2 to 5 verbatim lines per angle and scripts are scoped to a segment.

Scrapes, researches, ads and reports in one feed with type tabs and a date range. Competitor ad searches save to the account.

The app opens on a search box with a dark rail for navigation. Less chrome, faster to the first scrape.

find_early_products takes a brand filter and hints, and runs an evidence pass per product before it ranks them.