PaidMod is an API that reads the intent behind chat messages and stops buyers and sellers from moving the deal to Zelle, PayPal or WhatsApp, before your fee disappears.
Send us a sample of your chat logs. We show you what slipped through.
When two people meet on your platform and then pay each other elsewhere, you did the work of matching them and earned nothing. For a marketplace taking a percentage of each transaction, that is lost revenue on your best-matched deals.
Regex lists block “WhatsApp” and “Venmo”. People respond by typing “Z e l l e”, spelling out digits, or saying “message me on the other app”.
Whether the sender is trying to move payment or contact outside your platform, however they phrase it: spaced letters, spelled-out numbers, hints, emoji and slang.
Add one call to the chat pipeline you already have.
Your chat backend (Stream, Sendbird or custom) posts each message to PaidMod before delivery.
Two focused checks run in parallel: off-platform payment, and hidden contact details. Sender history adds context.
Block, mask, warn or send to audit. PaidMod recommends; your platform decides.
Structured, typed responses mean no parsing free text. Synchronous webhook or async stream inspection.
POST /v1/moderate{
"conversation_id": "conv_88310",
"sender_id": "user_seller_401",
"receiver_id": "user_buyer_109",
"message_text": "…send me $35 via Z e l l e…"
}
Response{
"flagged": true,
"confidence": 0.98,
"risk_score": 9.8,
"detections": {
"off_platform_payment": true,
"obfuscated_contact_info": true
},
"recommended_action": "BLOCK_AND_MASK",
"sanitized_text": "Hey, [Message hidden…]",
"latency_ms": 92
}
Our focus is small and mid-size platforms with roughly $100k–$5M in monthly transaction volume, built on Sharetribe, Bubble, or custom Node and Laravel stacks. Teams too small for a trust and safety department, but big enough that leakage hurts.
Simple monthly tiers by evaluated messages. If recovered fees do not cover the cost, we want to know.
Share a sample of chat logs and we will return a free 7-day leakage audit: what was caught, what a keyword filter would have missed, and an estimate of the fees at stake.