--- title: "Socialnib — FAQ" description: "Setup, privacy, the free allowance, billing, drafting quality, and X account safety — the questions people ask before installing Socialnib." source: https://socialnib.com/faq --- # Socialnib — FAQ Setup, privacy, the free allowance, billing, drafting quality, and X account safety — the questions people ask before and after installing Socialnib. Install it from the Chrome Web Store: https://chromewebstore.google.com/detail/socialnib/okalkgcmbehjpfmaclhifegmlbcgehah ## Getting started ### What do I have to set up? Install the extension and sign in with Google. That's it. There's no model to choose, no API key, and no card. The coral icon appears in X's reply toolbar and drafting starts working immediately. Your plan and allowance are also visible on the web at https://socialnib.com/account, but you never have to go there for drafting to work. ### Why does it need a Google sign-in at all? Because we run the model for you, and the sign-in is what ties your free allowance to you so one person can't spend everyone else's. We use Google purely as identity — we don't read your Gmail, Drive, contacts or anything else, and we can't post to X on your behalf. ### Which browsers work? Any Chromium browser with Manifest V3 support: Chrome, Edge, Brave, Arc, Opera. Safari and Firefox aren't supported today because the extension relies on Chrome's extension APIs. ### Does it work on twitter.com as well as x.com? Both. The content script matches x.com and twitter.com and runs nowhere else. ### Can I build it from source instead? Yes. The extension is MIT licensed — clone the repository, run the build, and load the dist folder unpacked from chrome://extensions with developer mode on. It still signs in to our API and uses the same allowance, because that's where the model runs and we pay for it. ## Using it ### Will it overwrite something I've already typed? Never. It only writes into a reply box that is completely empty, it re-reads the box immediately before writing, and it checks again before every step of the insert. If you start typing while a draft is landing, the write is abandoned rather than finished over the top of you. ### Can I turn the automatic drafting off? Yes. Switch off pre-drafting under Advanced and nothing runs in the background. You then draft on demand with ⌘⇧Y (Ctrl+⇧+Y on Windows) or by clicking the coral icon in the reply toolbar. ### How do I get a different reply? Click the coral icon. Every reply already written for that post is in one table with its tone, shape, text and character count. Arrow keys move, 1–9 jump to a row, ↵ inserts, Esc closes. Opening it costs nothing, because the batch is cached for that post. ### Can I change the keyboard shortcut? Yes — go to chrome://extensions/shortcuts and rebind it to whatever you like. ### Does it read the whole thread or just one post? The thread. It sends the post you're replying to, the chain leading up to it, and some of the replies already there, which is why it doesn't restate the post back at you or miss context from earlier replies. You can narrow how much context it uses, or turn it off, under Advanced. ### Does it post for me? No, and it never will. It writes into the composer and stops there. You read the reply and press the button. Automating engagement breaks X's rules and gets accounts suspended, so it isn't something we'd build. ## Quality and voice ### Won't it just sound like AI? That's the core design problem, and it's most of the work. Instead of returning the first thing a model says, Socialnib writes a batch of about 8 candidates across the tones that suit the post, then runs a second pass that scores them against what you said you're optimising for. You also set your voice once, which is what stops every reply reading like a template. ### What are house rules? Free-text constraints applied to every generation. Things like “never open with Great point”, “no em dashes”, “lowercase only”, or “never more than one emoji”. They're the fastest way to strip the obvious tells out of model output. ### What is the “your stack” field for? It's the only thing that lets a reply make a truthful first-person claim. Write down what you actually use and have built, and the model can say “I use this” about those things. Claims about anything not listed get thrown out rather than invented. ### Which model does it use? We choose per post, and we change it whenever a better one ships. That's deliberate: pinning a model name here would turn an implementation detail into a promise, and the whole reason there's no model picker is so you get the upgrade without doing anything. ### Are all 20 tones written every time? No. Writing 40 replies means nobody reads them. The model picks the handful of tones that actually suit the post and writes those in both shapes, which lands on about 8 — enough for a real choice, few enough to fit on screen. Any tone you want that wasn't in the batch is one click away. ### What happens if a generation fails? Transient failures are retried with backoff, truncated responses are salvaged for whatever complete replies they contain, and tones the model skipped are filled with one follow-up call. If the judging pass fails you still get the batch, just unranked. And if nothing usable comes back at all, the generation isn't counted against your allowance. ## Allowance and billing ### What does it cost? The first 100 generations are free and need no card. After that Pro is $8 a month for 1,000 generations a month. ### What counts as one generation? One batch of replies for one post. Reopening the table for a post you've already generated for is free, re-ranking replies you already have is free, and a generation that produces nothing usable isn't counted against you. ### Do you offer refunds? No. The free allowance is there so you can judge the product before paying for it — no card, no timer, no expiry — so we don't refund change of mind or unused time. Billing errors are separate: if you're charged twice or charged the wrong amount, we fix it. The refund policy explains the reasoning and the exceptions. ### Is the free allowance per day or in total? In total, once. You get 100 generations from the day you sign in. They don't refill and they don't expire, so spending them over a week or over three months comes to the same thing. Pro is the one that resets: 1,000 on the 1st of each month. ### What happens when I run out? New generations stop and the extension says why. Replies already written stay available because they're cached in your browser. Subscribing turns generation back on immediately. ### Where can I see how much allowance I have left? Two places, both showing the same number because both read it from the same server: the extension's Account panel, and your account page at https://socialnib.com/account. The web one also shows your renewal date and lets you subscribe, cancel, pause or resume — useful if you're on a machine without the extension installed. ### Will you charge me without asking? No. There's no card on file and no trial that converts on a timer. When the free generations run out, drafting simply stops until you decide to pay — the charge only ever happens because you went through checkout. ### How do I cancel, pause, or change my card? All of it from your account page at https://socialnib.com/account, or from the extension's Account panel — no email to us and no ticket. Cancelling keeps Pro running until the end of the period you paid for and stops the renewal. Pausing stops the billing and the drafting together until you resume it. Card changes and invoices open in our payment provider's portal from that same page, because they take the payment and we never handle card details. ## Privacy and safety ### What gets sent to your servers? To write a reply we send the post you're replying to, the conversation around it, the post's image if you've left that setting on, your X handle, and your voice settings to api.socialnib.com. That's what the model needs. It is not used to train models. ### What stays on my machine? Your generation history, your favourites, and your settings. They live in your browser's extension storage and we never receive them. Clear any of it, or all of it, from the settings page. ### What do you store about me? Your Google account's email, name, profile picture and account id, so the sign-in works and your allowance is yours. Plus a running count of how many generations you've used, which is how the free allowance is enforced. We don't keep the posts you replied to or the replies we wrote. ### Is using an AI reply tool against X's rules? Using a tool to help you write isn't the problem; automating engagement at scale is. Socialnib deliberately never posts for you and has no bulk mode. Read X's current rules, keep a human in the loop, and don't use it to spam. ### What permissions does the extension request, and why? Storage, for your settings and history on your device. ActiveTab, so the icon and shortcut can act on the tab you're looking at. Identity, for the Google sign-in. Network access is limited to our own API domain, and the content script only runs on x.com and twitter.com — it has no access to your other tabs or your browsing history. ### Can I delete my account? Yes. Email privacy@socialnib.com and we'll delete your account and everything attached to it. Uninstalling the extension removes everything stored locally straight away. ## Still stuck? Email support@socialnib.com and describe what you were doing when it broke. Including your browser and whether it fails on every post makes it much faster to fix. Other routes in: https://socialnib.com/contact