Getting Started: Upload Your First 10 Posts
The foundation of great AI-generated content is quality training data. Learn exactly what to upload — and what to avoid.
Go to your LinkedIn or X profile
Navigate to your profile and open your post history. You're looking for posts you wrote yourself — not reshares, not company announcements, not posts that heavily quoted someone else. Your original voice, your original words.
Tip: Filter by 'Top posts' if available. These tend to reflect your most natural writing — posts where you were engaged enough to write something real.
Select 10 posts that sound most like you
Look for variety in topic but consistency in voice. Don't select 10 posts all about the same subject — that limits how well the model understands your overall style. But do select posts where the writing feels unfiltered. Posts where you had something to say and just said it.
Avoid anything that starts with 'I'm excited to announce' or 'Key learnings from' — these are usually written in performance mode, not your natural voice.
Copy or export the post text
You can copy-paste directly from each post, or export them as a text file. Draftveil accepts: plain text (.txt), or direct paste into the training interface. You don't need to include engagement metrics, dates, or any metadata — just the text of each post.
If a post includes a link, you can include or exclude it. The model won't learn from the linked content, so it doesn't matter either way.
Tip: If you have fewer than 10 original posts, use what you have. Seven strong posts are better than ten that include posts you didn't really write in your voice.',
Open the Draftveil training interface
From your Draftveil dashboard, click 'Train voice model' in the left sidebar. You'll see the training interface with a paste area and a file upload option. Either method works — use whichever is easier given how you collected your posts.
Upload and start training
Paste or upload your posts and click 'Start training.' The model runs in the background — you'll see a progress indicator. Training typically completes in 1–3 minutes depending on the volume of content.
While it trains, set your platform preferences: LinkedIn, X, or both. Set your default post length: short (under 150 words), medium (150–350 words), or long (350+). These preferences tell the model where to target its output.
Tip: You'll be notified when training is complete. From that point, every post you generate uses your personalized voice model — not a generic AI output.