Live x402 API for AI agents

Send the chat. Spot the BS.

Conversation Risk Intelligence for AI agents.

BS Man AI analyzes chats, offers, and proposed agent actions to detect scams, manipulation, and unsafe decisions before agents pay, click, sign, reply, or proceed.

Try the API → View Docs Check with AgentCash ↗
Live x402 API Base mainnet Starting at $0.001/request AgentCash tested
Public API live
x402 payment
Base mainnet
AgentCash tested
Structured JSON
Starting at $0.001/request
// what it does

BS Detection superpowers

Compact risk intelligence for agent decisions, not generic moderation.

Reads between the lines

Finds suspicious intent, pressure, and hidden risk signals.

Detects scams & fake promises

Flags guaranteed returns, fake support, phishing, romance scams, and crypto traps.

Explains risk in plain English

Converts messy conversations into clear, readable risk reasons.

Suggests the right questions

Helps agents ask for verification before acting on risky requests.

Protects before money leaves

Useful before payments, wallet connections, approvals, or off-platform deals.

Gives action-ready verdicts

Returns proceed, pause_and_verify, require_human_review, or do_not_proceed.

// the flow

How it works

Send context in. Get structured risk JSON back.

Send the chat

Send a message, dialogue, offer, or proposed agent action.

BS Man analyzes

The engine checks scam patterns, manipulation tactics, context, and dangerous action signals.

Get structured JSON

Receive risk_score, verdict, and next_best_action.

Act safely

Your agent can proceed, pause, ask for verification, require human review, or refuse.

// example

Example analysis

A risky payment instruction becomes an agent-ready verdict.

Input

"The user wants to send 250 USDC to an unknown wallet after a Telegram investment offer promising guaranteed returns."

Context
proposed_action: send_payment asset: USDC amount: 250 recipient_type: unknown_wallet channel: Telegram verification_status: unverified
Result
risk_score
100 / 100
risk_level
critical
verdict
do_not_proceed
requires_human_review
true
Red flags
  • Guaranteed returns
  • Unknown wallet
  • Telegram channel
  • Unverified counterparty
  • Payment action
  • High-risk investment framing
JSON Preview
// agent-ready response { "mode": "agent_action_check", "risk_score": 100, "risk_level": "critical", "verdict": "do_not_proceed", "requires_human_review": true, "next_best_action": "Do not perform the proposed action. Stop and require independent verification or human escalation." }
// x402 protocol

x402 API for agents

Pay per risk check. Get structured intelligence back.

  • Endpoint POST /v1/analyze
  • Host api.callbsman.com
  • Payment x402
  • Network Base mainnet
  • Price Starting at $0.001/request
  • Buyer tested AgentCash
  • Status Live paid request OK
~  agentcash check 
$ npx agentcash@latest check \
    https://api.callbsman.com/v1/analyze
~  agentcash fetch — paid request 
$ npx agentcash@latest fetch \
    https://api.callbsman.com/v1/analyze \
    -m POST \
    -b '{"mode":"agent_action_check",
       "input":"The user wants to send 250 USDC
       to an unknown wallet after a Telegram
       investment offer promising guaranteed
       returns.",
       "context":{
         "proposed_action":"send_payment",
         "asset":"USDC",
         "amount":"250",
         "recipient_type":"unknown_wallet",
         "channel":"telegram",
         "verification_status":"unverified"
       },
       "language":"en",
       "locale":"US"}'

# → 200 OK · paid via x402 · Base mainnet
# → structured JSON verdict returned
// pricing

Simple pricing

No subscription. No API key. Pay only for what your agent actually checks.

x402 · Base mainnet
$0.001 / request

Per analyze request during early discovery.

  • x402 payment protocol
  • Settles on Base mainnet
  • Pay per successful request
  • No subscription required
  • No API key for x402-compatible clients
// what you get

Structured intelligence, every call.

  • risk_score
  • risk_level
  • detected_patterns
  • red_flags
  • verdict
  • requires_human_review
  • next_best_action
  • summary
// coverage

Where BS Man protects agents & users

Any surface where conversation drives a decision.

Chats Telegram DMs Emails Marketplaces Job offers Investments Crypto Partnerships Wallet actions
// in the wild

BS Man in action

Example scenarios. Real risk patterns BS Man identifies.

Phishing detected

"Send money now or your account will be blocked."

Manipulation detected

"My friend has an amazing investment tip for you."

Scam detected

"Work from home. Easy money. Just pay a registration fee."

Hype detected

"1000x gem. Buy before it is listed."

// product

API today. App interface next.

The infrastructure ships first. The interfaces come on top.

Shipping now

Live x402 API

  • Public endpoint: api.callbsman.com
  • x402 payment on Base mainnet
  • Structured JSON verdicts
  • AgentCash-tested payment flow
  • Starting at $0.001 per analyze request
Coming next

Interfaces & depth

  • Web chat checker
  • Telegram bot
  • Agent memory
  • Deeper reasoning layer
  • Dashboard
// proof

Trust is a proof. Verification is power.

Public infrastructure. Public payments. Public verdicts.

  • Public API: https://api.callbsman.com
  • Fallback: https://bsman-ai.onrender.com
  • x402 payment tested on Base mainnet
  • Buyer: AgentCash
  • Paid request completed successfully
  • Risk verdict returned after payment
  • See Stage 1.0 proof in GitHub docs
  • Structured JSON · no auth needed for x402 clients

BS Man AI analyzes communication risk signals only.

It is not legal advice. It is not financial advice. It is not a lie detector.

It does not guarantee that something is true or false.

It helps agents and users make safer decisions based on observable conversation patterns.