Draft LabLearn from posts you may use · publish what you approve

Sign in to your workspace

Paste the workspace token from WORKSPACE_TOKEN in the server's .env.local. It is stored as an 8-hour session cookie.

Draft Lab

Turn a reference post you have the right to use into four original drafts for your brand. You review them, approve one exact version, then publish it or add it to the schedule.

  1. Choose a brand
  2. Add the reference post
  3. Say what your post adds
  4. Generate four drafts
  5. Review and approve one
  6. Publish and measure

Brand

Drafts follow this brand's audience, voice and rules, and everything the app learns is kept per brand. Saved rules can't be edited; create a new brand to change them.

Reuse the same brand so learning accumulates over time.

Brand setup

Create the brand once. Rules are fixed after saving, so learning always refers to the same brand.

Propose a profile from your own posts

Optional: paste 1–10 posts you own (10,000 characters each). Model output is a hypothesis draft, not evidence. Nothing saves or changes the form automatically.

How you want to be known; editorial direction, never factual evidence.
Topics or phrases to review carefully, one per line (100 items, 2,000 characters each). Advisory; prohibited claims remain the separate hard gate.
Who the posts are for.
IANA name, e.g. America/Los_Angeles. Needed for weekly learning reviews.
X preferences

Optional tone, length and CTA habits. Short/medium/long upper preferences: X 100/200/280; LinkedIn 300/1,000/3,000 characters. These do not replace platform limits or human brand-fit review.

LinkedIn preferences

Optional tone, length and CTA habits. Short/medium/long upper preferences: X 100/200/280; LinkedIn 300/1,000/3,000 characters. These do not replace platform limits or human brand-fit review.

Name your recurring topics, one per line (up to 100 names of 200 characters). Tags help you inspect topic imbalance; they do not change learning.
What the posts should achieve, e.g. "grow a founder audience". One per line; up to 100 items of 2000 characters.
Sentences or short posts that sound like you. The writer imitates their tone. One per line; up to 100 items of 10000 characters.
Statements the brand must never make. Drafts containing them are blocked. One per line; up to 100 items of 2000 characters.
Calls to action you like to end with. One per line; up to 100 items of 2000 characters.
Empty lines are ignored. Saved brand rules stay fixed.

Reference post

The post whose structure you learn from: its hook, argument and evidence, never its wording. Import an X or LinkedIn post by URL, or paste the text.

An X or LinkedIn post. Its public engagement is shown as context, never as a score.
Paste the post here. Only use posts you have the right to use.
Required. A public post is not automatically yours to reuse.
Needed for analysis and writing with a live model. Leave off to keep the post away from external models.
Only for your own posts. Marks it as owned content that may be republished. Not available for imported posts.

What your post adds

Each draft must add something the reference doesn't have: your own procedure or first-party evidence. That keeps the post original, and every factual claim backed.

A procedure readers can follow, supported by the evidence below.
What your version does better, in at least 40 characters. Leave empty and the writer proposes one. A new sentence starts a new content family.

Supporting evidence

A verified fact the drafts may state. Drafts can only make factual claims backed by evidence like this.

Where it comes from.
When someone last checked it.

Generate four drafts

The app analyses the reference, then writes V0, an improved baseline, and three variations for you to compare.

Length and format rules follow the platform.
How the variations differExplore angles: V1–V3 take different angles on the same contribution.