Your real work is your best content.

Paje lives inside Claude. It watches your real conversations — strategy sessions, code reviews, planning calls — and turns them into ranked LinkedIn and X drafts in your voice. No blank prompts. No making things up.

Set up Paje Pricing

What Paje produces from a real work week

Throughline

“This week you restructured GTM around funnel stages instead of tools, committed to MQL→SQL as the single conversion metric, and cut reporting time from hours to automated.”

Draft #1

Reach 9Risk 2

“I flattened our GTM team this week. Four people, each owning a funnel stage end-to-end — not a tool. Nobody ‘owns AI.’ Everyone ships with it. Here's what broke and what got better immediately...”

Draft #2

Reach 8Risk 1

“We spent 6 hours a week building pipeline reports. Now it's automated and the team optimizes instead of collecting data...”

+ 3 more drafts, ranked by reach vs. risk. Confidential items held for review.

How it works

Install in one sentence

Tell your AI agent: "Install Paje MCP." It fetches the config from paje.ai and writes it. No terminal.

Work like normal

Use Claude for real work — strategy calls, code reviews, planning sessions. Paje captures the notable parts in the background.

Get your weekly brief

Say "weekly brief." Paje finds the throughline across your week and produces ranked LinkedIn drafts grounded in what you actually did.

Edit, publish, repeat

Review drafts on your dashboard. Edit in place, publish to LinkedIn or X with one click, or hold anything that needs more thought.

What makes it different

Confidentiality built in

Client names, credentials, internal metrics, and private IPs are scrubbed before anything is stored or sent to a model. Risky drafts are held for your review.

Your voice, not AI voice

Set your tone, audience, and anti-patterns once. Drafts come out first-person, specific, and grounded in your real decisions — not generic thought-leadership filler.

Ranked by reach vs. risk

Each draft gets a reach score and a risk score. High-reach, low-risk posts surface first. You see exactly why something is ranked where it is.

Set up in one sentence

Paste this into Claude Code, Cursor, or any agent that can edit files:

“Install the Paje MCP server into my Claude Desktop config. Get the config from paje.ai/api/mcp/config and set it up for me.”

Full setup guide

Questions

Will it post something I shouldn't share?: No. A multi-layer confidentiality filter runs locally on your machine before anything is stored. Client names, API keys, connection strings, and financial details are scrubbed or flagged. Drafts that reference sensitive material are held for your explicit review — they never auto-publish.
Does it just summarize my chats?: No. Summaries aren't posts. Paje extracts the specific decisions, things you shipped, and problems you solved — then writes them as posts in your voice with a hook, a body, and a closing. It's the difference between meeting notes and something people actually want to read.
What if my work isn't "interesting enough" for posts?: It is. Most people don't realize how much shareable material is in their day-to-day. Restructuring a team, choosing a vendor, debugging a production issue — those are real stories. Paje finds them because it sees your actual work, not a blank prompt.
Do I need an Anthropic API key?: Optional. With a key, Paje uses Claude to extract work items and write drafts (higher quality). Without one, it falls back to heuristics — still works, just less polished.

Stop starting from a blank prompt.

Your real work is already happening. Paje turns it into posts.