replies, written in your voice.
You just approve.

loom is an AI employee that learns your voice and writes on X for you. You add the sources (news sites / RSS, X accounts you follow). loom watches in the background — when a new article is published or someone you follow tweets, it drafts in your style. Every draft comes to you for approval: what you approve is posted in one tap or scheduled to your best time. Nothing is ever published without your approval.

You give a writer you admire or your own X handle — the system auto-scans recent tweets and learns the style. You don't hand over tweets one by one. When it writes, it's not just tonal similarity — your micro-habits like punctuation, line breaks, word choice and opening lines are preserved to the letter. On top of that you pick from 9 content types (hot take, story, list, question, build-in-public, etc.) and 6 tones (professional, casual, provocative, witty, thought-provoking, direct).

Control is entirely yours — nothing is published without your approval; that's a deliberate design decision, not a toggle. There are two modes: (1) Approval — loom watches your sources and drafts; every draft comes to you, and you say 'post / edit / schedule to best time / cancel'. (2) Manual — loom only produces when you assign a task. The daily production limit and working hours are under your control too.

Yes — every approval, edit or rejection is written to the system's memory. It constantly tracks which style you like, which topics you reject, which opening lines work; it uses those patterns in later drafts. In the command center, the 'what it learned today' card shows what it reinforced, and the decision log shows step by step which draft it produced from which source and why.

Connecting an X account is open on all plans — Free and Starter 1 account, Pro 3, Power 10. Don't want to connect? No problem: the system prepares drafts, you copy and post on X yourself. If you connect, one-tap quick posting is unlocked. Scheduling to a date or your best time is available on Pro and above (Pro 5/day, Power 30/day).

No, we can't — it's technically impossible. The connection uses X's official OAuth infrastructure: you enter your password on X's own page, and it never reaches tweetloom. The permission you grant covers reading + posting only; we can't access your DMs, change your account settings, or take over your account. You can revoke access anytime from your X settings (Settings → Security → Connected apps) with one click — you don't even need to ask us. And if you'd still rather not: you can use tweetloom without connecting at all — copy the drafts and post them yourself.

No — nothing is ever published without your approval. That's not a setting, it's the product's design: replies, quotes, tweets, threads — everything lands in the approval queue first; what you approve goes out in one tap or gets scheduled to your best time. tweetloom is designed to work with X's rules, not around them — the goal isn't to game the algorithm, it's to add content to X that's genuinely worth reading. Posts go through X's official API with the OAuth session you authorized (the same path Hootsuite and Buffer have used for years). Daily limits are human-scale and built into the product — there's simply no setting that reaches spam tempo. Engagement pods, automated DMs, follower inflation: none of these exist in the product. Anytime, "stop loom" turns it off entirely. Our founder runs the product on his own account — the rules apply to us the same way.

No card needed. The 7-day trial is completely card-free — since we never have your card details, a 'trial ended, money silently charged' situation is technically impossible. If you subscribe: cancelling is always one click, and you keep access until the end of the period. If you delete your account, your subscription is cancelled automatically — 'I deleted my account but kept getting charged' is structurally impossible here (deletion won't complete until the cancellation is confirmed). Credit numbers aren't inflated (1 credit = 1 tweet), and there's a 14-day money-back guarantee. Payments are processed by an independent payment provider — we never see your card details.

Simple: 1 credit = 1 tweet (medium length). Most tools charge 5-10 credits per tweet — we keep it 1:1, so instead of inflated numbers like '12,000 credits' you see how many tweets you actually produce. Credits only rise for long output: Tweet/Reply/Quote → Short & Medium 1, Long 2, Extra-long 4, Max 8 credits. Thread → short (3-5 tweets) 5, medium (5-7) 8, long (7-10) 12 credits. Credits are deducted when production completes; if there's a technical error they aren't. 'Regenerate' starts a new production and uses its own credit — which is why we keep the most-used lengths at 1 credit. Free comes with a 30-credit welcome bonus.

If you run out mid-month you don't need to switch plans: buy a one-time credit pack from your account — 200 credits $12 · 500 credits $25 · 1,500 credits $60 · 5,000 credits $160. Pack credits stack on top of your monthly credits and never expire.

Every tweet, reply, quote and thread is scored 0-100 automatically. 75 and up is ideal, 50-74 neutral, below 50 warns. The score doesn't tell you 'how many likes you'll get' — it checks the structural traits X's algorithm likes (strong opening, suitable length, hashtag count, issues like engagement-baiting). On a low score you can regenerate in one click.

Native in both — no templated feel, natural. Even if style training is done with an English writer, you choose the output language: you can borrow an English writer's rhythm for Turkish content, or the other way around. Proper nouns and grammar are always kept correct.

If you post a few tweets a week and just want a helper, Starter ($15) is enough. If you produce regularly and want to train multiple styles with more credits, Pro ($35). If you want an AI employee to run your account for you, Power ($59) — loom is included on Power, no extra charge.