💼 LinkedIn · 🐦 X · 👥 Facebook · 📸 Instagram · 👽 Reddit · ▶️ YouTube

What do they actually mean?

Decode the professional-speak, humble-bragging, and tech-bro hype of social media KOLs. Runs on your own AI key or a free local model. No account, no data collected - your posts never touch our servers.

Add to Chrome - 5 free decodes See how it works →
↓

Decodes are AI-generated satire and opinion - not statements of fact about any real person. Examples shown are generated with Claude (a cloud model); a free local model works too, with shorter, rougher output.

988K
people mock these posts on r/LinkedInLunatics. We just automate it.
6
platforms covered - not just LinkedIn
100%
of your data stays off our servers
How it works

Two minutes to set up. No account, ever.

Install the extension. Browse normally. Click to decode.

Step 01
🧩

Install and pick your AI

Add to Chrome in one click. Then either run a free local model (Ollama) or paste your own Claude or OpenAI key - a decode costs about a cent on your own key, $0 locally. 5 free decodes to try, no account.

Step 02
📜

Browse your feed normally

Decode buttons appear automatically on posts across LinkedIn, X, Facebook, Instagram, Reddit, and YouTube.

Step 03
💡

Pick your tone and read the truth

Choose Facts for the dry version, Sarcastic for wit, or Roast for maximum drama.

Two-way

It also works in reverse.

Write like a thought leader. Instantly. For any platform.

What you actually mean
"I got a new job."
↓
Also in the box

Three features that didn't fit on a bumper sticker.

Included with every license.

💀

Cringe Index

Every decode scores the post 0-100 on performative cringe, with a one-click share to X. "Peak LinkedIn" is a measurable quantity now.

🧹

Feed cleaner

Optional toggles hide promoted posts, "Suggested" cards, and "Recommended for you" clutter - locale-aware, off by default, one click to undo.

🎛️

Custom modes

Write your own decode or encode styles - "explain like a tired HR manager", "rewrite as a crypto bro" - and they appear as buttons everywhere.

Why Decoder

Six platforms. Your key or local. Actually fun.

The only tool that does all of this in one place.

Decoder BuzzKill linkedn't Corporate Speak Translator
Platforms ✓ All 6 LinkedIn only LinkedIn only LinkedIn only
Free local model (no per-use cost) ✓ Ollama - cloud key only - cloud only -
No account / sign-in needed ✓ ✓ - sign-in required ✓
Full-post decode with tones ✓ 3 tones + custom ✓ 3 tones ✓ - word highlights
Encode (plain text → influencer voice) ✓ per platform ✓ - -
Hide sponsored / feed clutter ✓ optional ✓ - -
Pricing $19 once - all 6 platforms, Encode, Cringe Index Free (your key) 30 free, then $4+ packs or your key Free
Pricing

One price. Pay once. Use forever.

Try free, then pay once. No subscription.

Launch price
One-time license
$19 once
No subscription. Pay once, use forever.
  • Unlimited decode & Encode
  • All 6 platforms
  • Bring your own Claude/OpenAI key, or run a free local model
  • Cringe Index - score any post 0-100 · Feed cleaner · Custom modes
  • 14-day money-back guarantee
Add to Chrome →

Start with 5 free decodes - no card needed.

FAQ

Questions.

You choose. Either paste your own Claude (Anthropic) or OpenAI key, or run a free local model with Ollama - in local mode nothing ever leaves your computer. Decoder charges a one-time license for the software; it never charges for AI usage and never sees your key.
No. It's a one-time $19 license - pay once, use forever. There are no monthly plans. Start with 5 free decodes, and if it's not for you there's a 14-day money-back guarantee.
The extension only processes text when you explicitly click a decode button - we don't scan your feed passively. Your post text goes directly from your browser to the AI provider you chose (with your key), or stays entirely local in Ollama mode. We run no server in that path and store nothing. Full details in the privacy policy.
Because each platform has its own performance style. LinkedIn is humble-bragging dressed as wisdom. X is tech-bro hype. Instagram is lifestyle branding. Facebook is MLM energy. Reddit is karma-farming. YouTube is like-and-subscribe filler. We have separate per-platform prompt tuning for all six so the decoded output actually makes sense for the platform you're on.
Facts is a dry, neutral restatement with zero editorializing - just what actually happened. Sarcastic names the performance with some wit and amusement. Roast is full commentator energy - maximum sarcasm, genuinely funny, still accurate. Same post, three very different reads.
Frighteningly well. Type "I got a new job" and watch it become a 5-paragraph LinkedIn post about your "next chapter" and the "lessons learned" on your "journey." Each platform gets its own voice - LinkedIn gets thought-leader posturing, Twitter gets indie-hacker metrics, Instagram gets a hashtag block.
Roughly a cent or less per decode on your own key (a decode is a few hundred tokens on a small model like Claude Haiku or GPT-4o-mini). In Ollama mode it's $0 - your own hardware does the work. Settings shows a running spend estimate so there are no surprises.
The decode buttons simply stop working and point you to the $19 one-time license. Nothing is deleted, nothing nags on a timer, and your settings stay put. Activate a license key and everything unlocks instantly.
Chrome today (Manifest V3). Edge and other Chromium browsers can typically load Chrome extensions too. Firefox and Safari are on the roadmap.
Decoder only reads the posts already on your screen and shows you a private translation - it never posts, likes, connects, messages, or automates anything on your account. The automation and scraping that get extensions in trouble are avoided by design. That said, every platform's terms are its own - see ours for details.
Install Ollama (free, ollama.com), pull a small instruct model like llama3.2 or qwen2.5:7b, and start it with OLLAMA_ORIGINS=chrome-extension://*. A 7B model runs comfortably on ~8GB of RAM/VRAM. Output is a bit shorter and rougher than the cloud models - free has a flavor.
No. You're just restoring signal. The Decoder doesn't make anyone look bad - it just removes the performance layer so you can read the actual information underneath. That's a public service.