Paste three posts
Old LinkedIn posts, emails or notes. Anything that sounds like you. That's the whole setup.
An AI that learns your voice and handles your LinkedIn.
Free forever tier.No credit card.Set up in sixty seconds.
Nothing leaves your browser. Paste any LinkedIn post to audit reach suppression flags in real time.
High Suppression Risk
suppressedFeed suppression triggers detected (external links, low variation, or AI markers). Address flags before publishing.
Sentence rhythm
BalancedReasonable spread. One short line after the longest would sharpen it.
What changed
It does not ban AI writing. It demotes generic writing.
LinkedIn replaced its ranking stack with a unified retrieval and ranking system, and its editorial team went public about suppressing low-effort AI content at the same time.
Templated openers, bullet stacks with no voice, flat sentence rhythm, engagement bait, links in the body: posts flagged this way stop travelling past your own network.
Meanwhile the signals that matter moved to dwell time, saves and real comments. Likes are close to worthless, and post-and-ghost is itself a negative signal.
How it works
No questionnaire about your brand. Your own writing does the teaching, and the model keeps learning from what lands.
Old LinkedIn posts, emails or notes. Anything that sounds like you. That's the whole setup.
Sentence length, how you open, punctuation, the phrases that are yours. Every number is visible.
After you publish, it retrains on your best posts, not the three you picked. Every version is kept and you can roll back.
Pick who you are and see how it opens a post for you. These are samples we wrote to show the range, not customer quotes.
Sample opener for a founder
We lost our biggest customer in March. It was the best thing that happened to the product.
Languages
Every tool in this category is English first. LinkedIn's professional base in MENA, DACH, France and South Asia is large, growing, and badly served.
Arabic done properly is unglamorous work that a San Francisco team will not prioritise.
خفّضت قيمة عقد أحد العملاء إلى النصف الربع الماضي. وارتفعت الإيرادات بنسبة ٣١٪ بحلول يونيو.
Right to left, with the connective structure Arabic professional writing actually uses.
What you get
A voice model per language, its own hook library, its own formatting habits. Arabic is right to left and reads like a person wrote it.
Trained on your best performing posts, not the three you picked by hand. The only version of voice training that improves instead of drifting.
A knowledge base of your real numbers, outcomes and opinions, pulled into generation. Style alone produces posts that sound like you and say nothing.
Need extra in a busy month? Add 10 posts for $5, 25 for $10 or 50 for $18 from inside the app. Top-ups sit on your plan and never expire.
We launched this year. Lots of tools in this space open with a logo wall and a 340% engagement claim. You can check ours instead: the reach check above runs for real, and we post our own results on LinkedIn as they come in.
Your founding price stays locked for as long as you're subscribed. You get direct access to us and a real say in what we build next.
Claim a founding placeFree preview launch week. Paid tiers and top-ups unlock next week!
Trying the voice and the Reach Check.
Solo creators posting a few times a week.
Founders and execs who want it on autopilot.
Agencies and ghostwriters running clients.
We never ask for your password, never touch your browser, and never use your drafts to train public AI.
You log in on LinkedIn's own page. We only get permission to post for you, and we never see your password.
Your draft is scored in your browser while you type. It isn't sent anywhere to be checked.
Your writing samples are encrypted and kept separate. They are never used to train public AI.
We publish through LinkedIn's API. No bots, no fake clicks, no risk to your account.
Disconnect LinkedKraft anytime with one click.
Questions
LinkedIn does not penalise AI writing as a category. It suppresses generic content. That distinction is the whole reason the Reach Risk Check exists: it checks your draft against the patterns the March 2026 feed update actually demotes, and tells you before you post rather than after.
Two things. Most voice models train once on three posts you picked and then never learn again. Ours retrains on the posts that actually performed, which is a different and better set. And most tools are English-first with translation bolted on. We build a separate voice model per language, including Arabic right to left.
Only when you press publish. We use LinkedIn's official publishing API for your own profile. We never automate connections, messages or engagement, and we never log into your account or scrape it.
Your samples train your voice model and nothing else. They are not used to train anything shared across accounts, and they are not sold. Delete your account and they go with it, properly, not flagged as hidden.
Not yet, and we are not going to invent any. We launched in 2026 and we are working with our first hundred users. What we can show you is the product running live at the top of this page, and our own results posted publicly as we get them.
You get told, with the number, and offered a top-up pack or a plan change. Nothing silently degrades and nothing gets slower.
Any time, from the billing page, in two clicks. You keep access until the end of the period you paid for, then the account drops to Free. Your posts, voice models and knowledge base stay where they are.
Five posts a month, free, no card. Long enough to find out if it actually sounds like you.
Free forever tier.No credit card.Set up in sixty seconds.