Threads Lab
coconect.io · @bronnikov.ca
ai: loading... scraper: loading...

Today at a glance

Loading…

Format performance (RL leaderboard)

higher score = the model will pick it more often
FormatPostsAvg scoreAvg viewsAvg repliesClicksSignups
No data yet. Generate posts and submit stats to populate.

How this works

  1. Scrape viral threads → go to "Viral Feed" tab, paste URLs or keywords. The system finds posts with high replies (Threads loves arguments) that sell something.
  2. Analyze the funnel → each viral post gets parsed for hook, format, trigger, and what it's selling.
  3. Generate 5 posts → "Generate" tab. Uses the viral patterns + what's been working for you.
  4. Post to @bronnikov.ca → copy each card, paste to Threads.
  5. Submit stats next day → "My Posts" tab. The model learns and picks better formats tomorrow.

🌱 Auto-discovery keywords

cron daily 06:00 UTC + the big button below to run them now
loading…

Runs every active keyword through the scraper. Needs Web Unlocker for keyword search (without it, you'll get a status message explaining what to do).

Add by URL

Search by keyword now

Paste manually

Viral feed

✏️ Polish my draft

paste your raw idea — AI rewrites it in your voice with viral patterns

🤖 Autonomous Bot Workflow

The bot autonomously researches, thinks, and generates posts. It learns from your feedback and analytics.

Status: SLEEPING Click 'Wake Up Bot' to start the daily flow.
😴Sleeping
🔬Researching
🧠Thinking
Ready
No posts generated yet. Wake up the bot.

📌 Pinned profile: @bronnikov.ca

scraped + studied for RL learning

One click pulls every visible post (text + views + likes + replies + reposts), saves them as your posts, and feeds the RL loop. Run this whenever you've posted new stuff on Threads.

My posts (latest 50)

🎙️ Voice profile

Hard rules the generator MUST follow. Built from analyzing your @bronnikov.ca posts + the feedback you give on each generated post. This is what makes the AI sound like YOU, not like AI.

📚 База знаний (Knowledge Base)

After you submit stats and click "extract insights" on a post, the model writes durable learnings here. These feed back into every future generation.