โ— coordination & authenticity engine ยท one API

Detect coordinated & fake activity.

Fake reviews, bot signups, engagement rings, purchased stars, sybil attacks โ€” they share one fingerprint: the same accounts acting together, fast. Verikast scores any (actor, target, time) data for that fingerprint. One endpoint. Any platform.

What it catches

One engine, any coordinated fraud

โญ

Fake stars & likes

Rings of accounts boosting the same repos/posts in tight bursts.

โœ๏ธ

Review farms

Paid reviewers hitting the same products together, out of pattern.

๐Ÿค–

Bot signups

Batch-created accounts with empty profiles acting in lockstep.

๐Ÿ“ฃ

Engagement pods

Coordinated like/reply rings inflating reach across platforms.

๐Ÿชช

Sybil attacks

One entity split across many wallets/accounts to fake breadth.

๐Ÿง‘โ€๐Ÿคโ€๐Ÿง‘

Vote brigading

Coordinated voting / brigades gaming governance & rankings.

Proof

It works โ€” measured, not claimed

Injected coordinated ring (fake-review scenario)18 / 18 caught
False positives on organic activity (200 real actors)0
Verdict when a ring dominates the dataINAUTHENTIC ยท 0/100
Open-social networks running live today3

Adversarially validated: it flags coordinated clusters while leaving genuine super-engagers untouched. A security screen, not a definitive verdict.

The API

One POST. A score + the ring.

Send engagement events โ€” who acted on what, and when. Add optional per-actor metadata (account age, followers, empty profile) to sharpen it. Bring data from any source โ€” reviews, sign-ups, on-chain, X/Twitter engagement you export.

# 1. get a free key (1000 checks/mo, self-serve) POST /v1/keys/new → { "api_key": "vk_โ€ฆ" } # 2. score any (actor, target, time) data POST https://verikast.com/v1/score x-api-key: vk_โ€ฆ { "events": [ {"actor":"u1", "target":"prod_A", "ts":1700000000}, // โ€ฆ ], "meta": { "u1": {"followers":1, "empty_profile":true} } }

The response pinpoints the coordinated accounts and clusters โ€” ready to block, flag, or review.

# response { "authenticity_score": 0, "verdict": "INAUTHENTIC", "quality_label": "Likely fake / coordinated", "explanation": "1 cluster โ€” 4 accounts acting together on shared targets.", "coordination_pct": 75.0, "flagged_count": 4, "flagged_actors": ["R0","R1","R2","R3"] }

Agent-ready: discoverable & callable over /mcp, with per-call payment via x402 for autonomous agents.

Pricing

Start free. Pay as you scale.

Free
$0
1,000 checks / month
  • Full detection engine
  • Score + flagged actors
  • Community support
Pro
$29 /mo
50,000 checks / month
  • Everything in Free
  • Higher rate limits
  • Batch endpoint
  • Priority support
Agents ยท x402
~$0.002
per check, pay-per-call
  • No subscription
  • Machine-payable
  • MCP-discoverable
  • Scales to zero

Infra runs lean โ€” the free tier exists to prove it on your data. Pricing illustrative during beta.

Verikast ยท a general coordination & authenticity engine ยท account check ยท Deep Audit ยท @verikast
A screen that surfaces coordinated patterns โ€” you decide what to do with them.