Skip to main content

​
Content Generation

One idea. Six platforms. No rewrites. DraftLift generates content by combining your memories, reference materials, and platform vessels with AI models. The result is content that sounds like you wrote it for the specific platform — not generic AI output pasted across channels. You don’t need to be a prompt engineer. You need an idea and something to say about it. DraftLift handles the rest.

​
The generation pipeline

Content creation follows a three-step pipeline:

​
Step 1: Select reference materials

Choose source materials that provide factual grounding. Reference materials are optional but improve output quality when you have specific data, articles, or documents to draw from.
  • Uploaded documents — PDFs and text files
  • Web URLs — Automatically extracted and summarized
  • Text snippets — Pasted content you’ve saved
When references are provided, the AI treats them as primary sources. If you didn’t provide it, it doesn’t get invented.

​
Step 2: Choose your vessel

Select the platform and content format. Each vessel has built-in rules for structure, length, tone, and platform-specific conventions: DraftLift also includes specialized short-form video vessels — Math Breakdown, Myth Buster, Aspirational Bridge, Open Loop Stack, and more — each designed for a specific storytelling format.

​
Step 3: Write your direction

Enter your topic or direction. This is where you tell DraftLift what to say — not how to format it. The vessel handles formatting. Your memories handle voice.
“Cold outreach on LinkedIn is dying because buyers have learned to ignore templated connection requests. The companies getting replies are the ones whose founders publish content that makes them recognizable before the first DM.”
Focus on the thesis. Be specific about your angle. For more on this, see Crafting Your Direction. Select your AI model (Pro+ tiers), then click Generate.

​
AI models

Starter tier users are assigned a single model. Pro+ tiers choose their preferred model per generation.

​
How context is assembled

Every generation combines three layers into a single prompt:
  1. Your memories — Active memories are included automatically. Platform-tagged memories are filtered to match the selected vessel (your LinkedIn-specific tone doesn’t bleed into casual X posts).
  2. Reference materials — Selected documents, URLs, and text provide factual grounding. Source URLs are linked naturally within the text when the vessel supports citations.
  3. Vessel rules — Format constraints, character limits, structural conventions, and platform best practices shape the output to fit the target channel.
This is why DraftLift output sounds different from generic AI. Your voice, your facts, and your style are baked in from the start.

​
Async generation

Content is generated asynchronously. Queue multiple pieces and navigate away — come back to your finished drafts when they’re ready. Real-time progress tracking shows the status of each generation.

​
Anti-slop quality checks

Before you see the output, DraftLift runs it through the Anti-Slop Quality System:
  • Critical patterns — Robotic phrases, AI vocabulary crutches, infomercial hooks
  • Warning patterns — Em-dash overuse, arrow symbols, formulaic triple-item lists
  • Style signals — Filler word density, adverb overuse, vague attributions
Flagged patterns are highlighted in the editor so you can fix them before publishing.