Most AI tools write blind.
We track every platform's algorithm so you don't have to.

For every platform we publish to, we keep a structured profile in plain language and structured weights: which formats earn the most reach this week, length sweet spots, hashtag rules, link rules, time-of-week weights, and the engagement signals each algorithm rewards heaviest. Our research team revises these as platforms ship changes; our daily canary flags shifts our research missed.

Every generation reads this profile and bends the LLM toward what the platform actually rewards. That is why the same prompt produces a different post for LinkedIn than it does for X, even before your brand voice gets layered on top.

See the live profiles on our public algorithm research feed.

You have two reels at 8 percent engagement and thirty regular posts at 4 percent. A naive recommender says “always post reels.” We don't. We use Bayesian credibility shrinkage so a small sample stays anchored to the platform's prior. Once you have enough data, your own numbers take over.

• Starter: see the recommended format as a chip in the content generator.
• Pro: opt-in toggle on campaigns to apply the recommendation across the whole batch.
• Agency: auto-apply by default, with one-click override per slot.

Industry Gold Slots are the candidate pool. Your own engagement history is the re-ranker. Once you have a couple of weeks of data, we keep a 168-cell map of your hour-of-week performance and combine it with the platform's own time weights. The result is recommendations like “Tuesday 8:17 PM beats your usual Saturday-morning slot for Instagram” — backed by your data, not someone else's.

Every platform has a first-hour cliff: low engagement in the first 15 to 60 minutes and the algorithm starves the post forever. When a post is stalled below your own median and the critical signals (comments, shares) are missing, our boost engine drafts a short voice-matched follow-up to restart the conveyor belt.

Off by default. Daily cap you control. Skipped automatically if the post is already getting comments. Every decision is logged so you can audit the ROI in your dashboard.

Every night we compare aggregate engagement across our user base on each platform against a 28-day baseline. When a meaningful shift shows up (down 25 percent on LinkedIn after a feed-ranking change, for example) you see a soft banner at the top of your dashboard with the platform, the magnitude, and what we are updating in your generations. Most teams hear about algorithm changes from newsletters; you hear about them from your dashboard.

Every layer is tier-gated and dual-write. Free and Starter users see the same byte-for-byte output as before this engine existed unless they opted in. Golden snapshot tests pin that behavior on every commit. If a platform profile turns out to be wrong, we can roll back to the previous version without touching your account or your scheduled posts.