How Bangerkit works: every feature

Problem: A new tool usually starts with an empty dashboard and manual set-up.

• You sign in on X's own page (OAuth); Bangerkit never sees your password.
• Once a plan, trial or gift code is active, a first import runs at no cost: up to 100 of your posts from the last 28 days with their edit history and current numbers through the official X API, your follower list through TwitterAPI.io, and the replies to you from the last 48 hours.
• Follows and unfollows are recorded from that day on, because X keeps no history of them.
• A short set-up asks for your app and post languages and a daily goal, suggests automations to switch on, and ends with a tour.

Plans: Every plan; the import starts once your plan, trial or gift code is active.

Problem: Knowing what to do first on X for an account of your size.

• Seven steps: install the Chrome extension, say who you talk to, post a "let's connect" post, turn on Reply Assistant, watch your followers, reply to people who want to connect, and set your daily goal.
• Each step ticks itself when your account data shows it done; none is ticked by hand.
• The advice and post templates follow your follower count: under 500, 500 to 1,999, 2,000 to 9,999, and 10,000 or more.
• A step that needs a higher plan is marked with the plan that includes it.

Plans: Every plan.

Problem: A tool alone doesn't say how to grow by talking to people.

• The Playbook tab shows the 21 pages of the Talk First playbook as images in the app; there is no PDF to download.
• A paid plan opens every page; during a trial or with a gift code you see the cover and the contents.

Plans: Every paid plan; a trial or gift code shows the cover and the contents.

Problem: Checking every tab to find out what needs doing today.

• Today's replies and posts count against a target, the larger of your usual day over the last two weeks and the daily goal you set (5 replies and 1 post until you have either); new followers, views and regulars answered can have goals too, and a streak counts the days you reach at least half the target.
• The next task and an Also waiting board point at what to do now, such as a post taking off with unanswered replies, a regular waiting, a product question, comments to answer, new followers or fresh posts in your Engage feed.
• Answer your people lists your posts of the last 3 days with people waiting for you (up to 6 posts), your regulars first, and you reply in a pop-up.
• On Creator and above, Ready to post shows the drafts Studio wrote for today.
• What should I do next? asks AI (OpenAI) for your next moves from today's numbers and what is waiting.
• While Home is open, the numbers of your posts from the last 48 hours are refreshed when you arrive and every 10 minutes, at no cost to you.

Plans: Every plan; Ready to post needs Studio (Creator and above).

Problem: Replies to you on X are spread across notifications, threads and old posts.

• The Replies tab collects comments under your posts and replies to you in conversations you joined, read through TwitterAPI.io with a 48-hour look-back.
• It checks every minute while the app is open on Pro and Power, every 5 minutes on Starter and Creator, and every 30 minutes on any plan while the app is closed, up to three times that when nothing new arrives, as long as you used the app in the last 7 days; after a longer break your next visit catches up on the last 48 hours.
• Each comment is Waiting, a Draft, Replied or Ignored; work post by post or through all waiting replies, sorted by newest, oldest, priority or most followers, with a card about each person: whether you follow each other, their followers, and how often you answered them.
• Reply publishes your answer through the official X API for 1 action, plus 1 per image; Open in X opens X with your text instead, for free.
• Starter takes in up to 2,500 comments created in any 30 days; the other plans have no such limit.

Plans: Every plan. New-reply checks while the app is open: Starter every 5 to 15 minutes, Creator every 5 to 15 minutes, Pro every 1 to 3 minutes, Power every 1 to 3 minutes. Comments the inbox takes in per 30 days: Starter 2,500, Creator no limit, Pro no limit, Power no limit.

Problem: Writing every reply from scratch takes time.

• Suggest reply with AI writes a draft from the saved conversation, and Rewrite draft with AI reworks yours in the conversation's language; both use OpenAI and only fill the reply box.
• TypeSafe AI checks each AI draft and shows how much it reads like generic AI writing (AI slop %).
• Always improve with AI, off by default, turns Reply into Improve & reply: AI polishes your text and sends its version without showing it to you again.
• Select up to 20 waiting comments and Draft with AI writes a draft for each; nothing is posted, and the drafts can go to the send queue.
• AI drafts cost no actions; only publishing a reply does.

Plans: Every plan. AI reply suggestions in any 24 hours: Starter 100, Creator no limit, Pro no limit, Power no limit.

Problem: Many comments are reactions or bots that don't need an answer.

• With comment recommendations allowed in Privacy settings, TypeSafe AI reads new unanswered comments of the last 3 days and sorts them: needs you, a simple reaction, an answer to your question, an opinion, or likely from a bot.
• Comments judged as needing no answer, or as likely from a bot, gather under Look finished, where you review them or skip them all.
• A suspected bot gets a Bot? badge; Mark as bot sends all of that account's comments, now and later, to Ignored.
• Sorting changes nothing on X and costs no actions.

Plans: Every plan; needs comment recommendations allowed in Privacy settings.

Problem: Quotes and tags are easy to miss in X's notifications.

• The Mentions tab lists posts that quote your posts or tag you, from the first check on; older ones are not fetched.
• It checks when you open the app, every 30 minutes while it is open and every hour while it is closed, through TwitterAPI.io.
• With comment recommendations allowed, TypeSafe AI marks each mention as important or simple.
• Write with AI and Polish with AI help with the answer; Send reply publishes it for 1 action, and Post on X instead opens X with your text for free.

Plans: Every plan.

Problem: Answering the same question again and again under one post.

• OpenAI sorts the waiting comments and mentions of the 48-hour inbox, once they are 5 minutes old, into questions, discussion, complaints and noise, and groups the questions under one post into topics.
• Sorting runs once 20 items are waiting or the oldest has waited 10 minutes, for up to 1,000 items in any 24 hours.
• You type one answer (up to 2,000 characters, in any language, with a product link if you like), and Reply Assistant writes each person's reply in their language from it.
• These replies follow Reply Assistant's mode and daily limit, a reply with a link always waits for your approval, and each published reply costs 1 action.
• Complaints are listed under Needs you personally, and noise can be skipped in one go.

Plans: Every plan; needs the Reply Assistant terms accepted and comment recommendations allowed.

Problem: Thanks, congratulations and short reactions pile up, and each one deserves an answer.

• TypeSafe AI sorts every new comment and mention into important, simple and other, and Reply Assistant writes replies in your voice with OpenAI to the simple ones only, never to important ones.
• Switches add comments that answer a question from your post, opinions about your post, mentions, and people who don't follow you: their other comments get a reply too, ending with an invitation to connect and always waiting for your approval.
• A draft the reviewer rejects is held back; one person gets at most 3 replies a day and one per conversation, and no wording is used more than twice in 24 hours.
• By default every reply waits as a draft; approved ones are published from your X account through the official X API, a few seconds apart, and Stop returns the unsent ones to your drafts.
• On Creator, Pro and Power, Send on its own publishes replies without asking you first; a reply with a link always waits for your approval.
• Each published reply costs 1 action, like one you send yourself; turning it on needs comment recommendations allowed and your consent to its terms, which say that X may limit accounts that use automation.

Plans: Every plan; on Starter it only drafts. Reply Assistant replies in any 24 hours: Starter 20, Creator 20, Pro 100, Power 300. Reply Assistant can send on its own: Starter no, Creator yes, Pro yes, Power yes.

Problem: Finding fresh posts worth replying to in your niche is slow.

• Type a topic or save it as a feed; Engage searches recent original posts on X through TwitterAPI.io, 20 posts a page, leaving out replies and reposts.
• Filters: people (everyone, your followers, people you follow, or new people who neither follow you nor are followed by you), language (English, Polish, German, Spanish or any), age (24 hours, 3 days or 7 days), up to 12 excluded words, and quote posts on or off; sort by latest or most discussed and hide posts you replied to.
• TypeSafe AI hides posts that are off-topic for you, such as events, token promotions and giveaways, and Also try suggests related searches.
• Write your reply beside the post: Reply on X opens X with your text for free, Add to queue keeps it for the Chrome extension, and Queue let's connect writes a short connect reply for up to 20 posts at once.
• Searches use the plan's daily searches, never actions; Search further reads older posts and uses one more search.

Plans: Every plan. Engage searches in any 24 hours: Starter 5, Creator 5, Pro 30, Power 90. Pages one Engage search reads (20 posts each): Starter up to 3, Creator up to 10, Pro up to 10, Power up to 10.

Problem: Remembering to search the same topics every few hours.

• Switch on auto-search for a saved feed and pick every 1, 3, 6, 12 or 24 hours; it runs only while the app is open.
• Each run uses one of the plan's daily searches, always leaves the last one for you, and reads up to 3 pages.
• Posts the AI judges to be invitations to connect get a short connect reply written by AI in your send queue; nothing is posted until you post it yourself.
• A run is skipped while 50 of its replies wait in the queue or the plan's daily Engage replies are already queued.

Plans: Every plan. Engage auto-search replies queued in any 24 hours: Starter 20, Creator 20, Pro 50, Power 150.

Problem: Keeping up with the people who keep showing up for you.

• Home lists posts from the last 24 hours by your top 20 repliers of the last 30 days.
• The list is read through TwitterAPI.io when you open Home, at most once every 10 minutes, and uses neither your daily Engage searches nor actions.
• Reply in a pop-up with Write with AI or Polish with AI, then Post on X opens X with your text.

Plans: Every plan.

Problem: Turning an idea into a post that doesn't sound generic.

• The Post button opens New post: up to 280 weighted characters, up to 4 images (JPEG, PNG or WebP, up to 5 MB each) and emoji.
• Write post with AI drafts from your saved posts and Improve post with AI reworks yours, through OpenAI; TypeSafe AI then shows how much the text reads like generic AI writing (AI slop %).
• Publish posts it from your X account through the official X API for 1 action, plus 1 per image; Schedule picks a later time instead.

Plans: Every plan.

Problem: Posting at the right time without being online then.

• Schedule a post from the composer or a Studio draft 2 minutes to 90 days ahead; up to 100 posts can wait at once.
• At its time Bangerkit publishes it from your X account through the official X API and charges its actions then, as Publish does.
• A post that can't go out within 30 minutes of its time is marked missed and never published late, and one whose outcome is unclear is never sent again on its own.
• The Queue tab lists upcoming and finished scheduled posts, where you can edit or cancel them.

Plans: Every plan.

Problem: A blank page every day, or posts that sound like everyone else's.

• Tell Studio what happened, typed or spoken (up to 30 minutes of recordings a day; OpenAI transcribes them and they are never stored), and it keeps the facts; About you and your own account data add more.
• Role models are up to 5 X accounts whose most viewed posts of the last 30 days teach it topics, length, hooks and structure, never their wording; they are read again every week for free.
• Write this day's posts, Write more and Describe posts write drafts with OpenAI, with a news step that searches the web; TypeSafe AI checks each draft, one that fails is rewritten once, and flags mark drafts that sound AI-written, aren't backed by your facts or will likely reach fewer people than usual.
• With Write for me on, Studio writes the day's posts on your first visit of the day and up to 2 more plans when a topic is hot in your niche; the drafts appear on Home under Ready to post. After 3 of its plans in a row go unused, it waits until you post, schedule or skip a draft or ask for posts.
• Nothing is published until you schedule a draft, and Results compares the first 48 hours of each scheduled Studio post with your usual reach.

Plans: Creator, Pro and Power; not on Starter. Up to 3 plans a day on every plan that has Studio. Studio drafts per day (UTC): Starter none, Creator 6, Pro 12, Power 24. Studio role-model reads you start in any 24 hours: Starter none, Creator 5, Pro 15, Power 30.

Problem: Posting many prepared replies and messages by hand without spending actions.

• Add replies from Engage, Home, Replies or Mentions to the Queue, and write direct messages there with Write with AI or Polish with AI; a message reaches people who follow you or accept message requests.
• Up to 200 items can wait, and up to 20 direct messages can be added in any 24 hours.
• The Chrome extension opens each item on X with your text filled in, and you press Post or Send yourself; posting this way costs no actions and uses no X API.
• After you mark a reply sent, the Queue looks for it on X for up to 2 hours and shows On X, or Not seen on X with Queue again.

Plans: Every plan.

Problem: Posting prepared replies one by one without copying and pasting.

• Install it from the Chrome Web Store and connect it in Queue with Connect this browser; it shows your send queue in Chrome's side panel.
• Open puts a reply or message into X's own composer through X's documented links with your text filled in; you press Post or Send yourself, then Sent → next.
• Start series opens the next reply by itself once X confirms it created the one you posted; a message always waits for Sent → next, and the series pauses when another X account is signed in.
• Follow before replying, off by default, has Bangerkit follow each reply's recipient through the official X API before the reply opens, for 1 action per follow, optionally only verified accounts.
• On x.com it reads only which X account is signed in, the address of the active tab and, for the reply you opened, whether X created it; it never clicks, types or posts on X, and never reads your reply text, headers, other cookies or other requests.

Plans: Every plan.

Problem: Seeing what works for another account before you learn from it or reply to it.

• Open any X profile or post and press Analyze in the extension's Profile view.
• Bangerkit reads that account's public profile and its own posts of the last 30 days (up to 200) through TwitterAPI.io, never through your X session.
• You get totals, median post numbers, posts per day, the best weekdays and hours in your time zone, the format mix and the top 6 posts.
• An account anyone analyzed in the last 24 hours opens from that read and doesn't count toward your limit; analyses cost no actions.

Plans: Every plan. Profile analyses in any 24 hours: Starter 3, Creator 10, Pro 30, Power 100.

Problem: X shows your follower count change, never who left.

• The follower watch reads your followers through TwitterAPI.io. On Pro and Power: the count every 10 minutes, a full comparison when it drops, and a full comparison every 6 hours anyway (every 12 hours from 10,000 followers, once a day from 50,000); your following list is read once a day.
• An account counts as gone once two comparisons in a row miss it, so an unfollow usually shows within half an hour on Pro and Power.
• Creator runs at half that pace: the count every 20 minutes, so an unfollow usually shows within an hour, and the full comparison every 12 hours (every 24 hours from 10,000 followers, every 2 days from 50,000).
• Followers shows your followers, following and not-following-back counts and a dated list of follows and unfollows; the history starts the day the watch starts, because X keeps none.
• On Starter the count is checked every hour, a drop starts a comparison no sooner than 12 hours after the last, the full comparison runs every 2 days and the following list once a day, and the watch pauses above 1,000 followers.
• The watch costs no actions; a follow or unfollow you click costs 1.

Plans: Every plan; on Starter for accounts with up to 1,000 followers. Follower count checks: Starter every hour, Creator every 20 minutes, Pro every 10 minutes, Power every 10 minutes. Followers the watch covers: Starter up to 1,000, Creator no limit, Pro no limit, Power no limit.

Problem: Following back everyone you talk to by hand.

• Auto follow back follows each new follower 30 to 120 seconds after the follower watch sees them.
• Auto unfollow unfollows accounts you follow that stop following you, 2 to 5 minutes after the watch sees it; more than 25 at once wait for your confirmation.
• Both go through the official X API, cost 1 action each, charged only when X confirms, and stay within the plan's daily limit and 40 follow changes in any 15 minutes.
• Both are off until you switch them on and confirm; X counts this as automated following, and you are responsible for it on your account.

Plans: Pro and Power. Automatic follows and unfollows in any 24 hours: Starter none, Creator none, Pro 50, Power 150.

Problem: A following list full of accounts that never followed back.

• Review lists the accounts you have followed for 48 hours or more that don't follow you, from the lists the follower watch saved.
• You unfollow them one at a time, and X is asked first whether they follow you now, so a new follower is never unfollowed.
• Each unfollow costs 1 action, charged only when X confirms, and at most 100 unfollows go out in any 24 hours, automatic ones included.

Plans: Every plan.

Problem: Knowing what to post more of without guessing.

• Overview compares 7, 14 or 28 days with the period before: net followers, impressions, engagement rate, posts and replies sent, with a posting calendar and streaks; Include replies adds your replies' impressions, as X's own analytics does.
• Content judges every post over its first 48 hours against your follower count (posts that pass are breakouts), shows younger posts racing toward that mark, and maps median views by weekday and three-hour window once a window has 5 finished posts on 3 days.
• Audience shows your followers and following, who replied and who came back, the comments you answered, everyone who talks to you, and What brings you followers: the posts after which commenters followed you.
• Share your numbers draws an image of the range in your browser; nothing is uploaded.
• The numbers come from the first import through the official X API, the inbox checks and Home's refresh through TwitterAPI.io, and Read current metrics in Studio, which reads up to 13 posts through the X API; reading analytics costs no actions.

Plans: Every plan.

Problem: Using your own AI assistant with your X data without copying it around.

• The MCP tab gives a server URL and a ready configuration for your MCP client; desktop clients connect through mcp-remote, which needs Node.js.
• Reading your posts, analytics, replies, Engage and the send queue is always on; you choose whether the agent may also create reply drafts, ignore comments, rewrite or skip Reply Assistant drafts and answer Questions, add replies to the send queue, run Engage searches (each uses a daily search) and add posts to Studio (up to 48 a day).
• It offers 24 tools, and the agent never publishes, hides, deletes or reports anything on X: approving and posting stay your clicks.
• Access has an expiry date and can be reset at any time.

Plans: Every plan.

Problem: Knowing what you pay for and what is left this month.

• Billing switches between monthly and annual billing (a year costs 10 months), starts the 3-day trial, and opens the payment portal for your card and invoices.
• Upgrade now charges you today, restarts your billing month and carries your unused actions over; a switch to another plan can also start with your next billing period.
• A one-time pack adds 1,000 actions for $29 to the current paid month; packs open once a trial has become a paid plan.
• Usage shows how many actions are left, when the pool renews and what each action costs; a gift code is redeemed in Billing.
• Cancel from Billing at any time; your plan stays until the paid period ends.

Plans: Every plan.

Problem: Deciding what is analyzed beyond what you ask for.

• Comment recommendations lets external AI providers analyze your saved comments and their conversations automatically; Reply Assistant, Questions and AI comment sorting need it.
• Product analytics sends limited usage events and a masked session replay to help improve the app; it is independent of the first choice.
• Nothing is analyzed or sent under either choice until you allow it, and you can change both at any time.

Plans: Every plan.

Problem: Working in one language while posting in another.

• App language sets the language of menus, messages and the AI's explanations and coaching: English, German, Spanish or Polish.
• Post language sets the language your new posts are written in; existing drafts and replies keep their own, and AI replies follow the language of the conversation.

Plans: Every plan.

Problem: Seeing what works for an account before you take a lesson from it.

• Type any public X account on the home page; the check reads its public profile and its own posts of the last 14 days through TwitterAPI.io.
• It works out how often the account posts, what a typical post earns and when its posts do best, and OpenAI writes a short read of what works and what you can borrow.
• One new check per visitor every 30 days; an account anyone checked in the last 30 days opens instantly, and results are not published.

Plans: Free, no account needed.

Problem: Spotting posts that read like generic AI writing.

• Enter an X handle; the check reads up to 20 of the account's own latest posts from the last 12 months through TwitterAPI.io.
• OpenAI scores each post 0 to 100 for the patterns AI writing overuses (template hooks, "it's not X, it's Y" turns, stock closers, lists of three, no specifics), and the account's score is the average.
• One new check per visitor in any 24 hours; an account anyone checked in the last 24 hours opens instantly, and a check is kept for 7 days so a shared link still opens.

Plans: Free, no account needed.