Proof, not points.
The KOL alternative for Web3.

Web3 projects spend $5,000 to $50,000 per KOL post for engagement that often comes from bot networks, never converts, and produces no defensible report. The audience belongs to the influencer — not to the project. The next launch starts from zero.

The pattern is familiar: a budget cycle gets approved, an agency books five KOLs, posts go out, the impressions look fine on the dashboard, and three weeks later the holders are gone and the project is asking why the numbers don’t convert.

The answer is that rented reach was never the same thing as activated community. Reach is rented. Community is owned. Most Web3 marketing budgets buy the first and call it the second.

What CommunityOS does instead

• Scans the project’s existing X follower base.
• Removes farmers and bots through the Bot-Kill filter before anyone reaches a queue.
• Scores remaining followers as Champion, Amplifier, Builder, or Early Adopter.
• Surfaces the Act Now queue — the specific people whose voice or reach would actually move the project this week.
• Tracks proof of every activation and pays only on verified action.

The result is a community report built on people who already chose the project — not strangers paid to perform interest.

Quest platforms made a clean assumption: if you reward an action, the action happens. The assumption was correct. What they didn’t account for was selection — the population willing to repeat a quest for $2 is not the population that will hold your token at $200.

The clickers got airdrop training. They built dedicated wallets, multi-account stacks, completion scripts. When the rewards stopped, they left. When the next project launched, they showed up there too. The graph of who actually held tokens after 90 days bore no resemblance to the graph of who completed the quests.

This is what CommunityOS replaces. Not better quests. Not better verification. A different shape entirely: archetype activation. Find the four people in your audience who would post a 12-tweet thread without being asked. Activate them. Verify the work. Pay them. Repeat with the next four.

The volume is lower. The conviction is higher. The cost per real holder is a fraction of the quest-platform model.

Mintlayer is the first Web3 project to run a full CommunityOS scan in production. The numbers, with permission:

• 5,806 real followers ranked after Bot-Kill
• 298 Act Now candidates surfaced in the first scan
• 1,440 Act Soft (warming up, not yet)
• 3,342 Monitor (the long tail)
• 25 Ignore (after Bot-Kill, residual low-signal accounts)

The full case study, including the agency cascade that opened from this single scan, is at /case-studies/mintlayer.