Documentation

Everything you need to integrate AIS into your agent.

Quick Start

Get protected in under 2 minutes.

Register your agent

curl -X POST https://ais.solpay.cash/api/register \
  -H "Content-Type: application/json" \
  -d '{"agent_name": "my-agent"}'

Save the api_key from the response. It won't be shown again.

Route requests through AIS

curl -X POST https://ais.solpay.cash/api/proxy \
  -H "Authorization: Bearer ais_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "tool": "web_fetch",
    "target_url": "https://api.example.com/data",
    "method": "GET"
  }'

That's it

Every request is now scanned for threats, secrets are automatically redacted, and attack signatures are shared with the network.

API Reference

Base URL: https://ais.solpay.cash/api

POST/register

Request Body

{
  "agent_name": "my-agent",
  "wallet_address": "optional-solana-address"
}

Response

{
  "ok": true,
  "agent": {
    "id": 1,
    "name": "my-agent",
    "tier": "free"
  },
  "api_key": "ais_abc123...",
  "proxy_url": "https://ais.solpay.cash/api/proxy"
}

POST/proxy

Proxy a tool call through AIS

Headers

Authorization: Bearer ais_your_api_key

Request Body

{
  "tool": "web_fetch",
  "target_url": "https://api.example.com",
  "method": "GET",
  "data": {},
  "headers": {}
}

Success Response

{
  "ok": true,
  "decision": "allow",
  "proxy_response": { ... },
  "secrets_redacted": 0,
  "latency_ms": 45
}

Threat Blocked Response

{
  "ok": false,
  "error": "Threat detected and blocked",
  "code": "threat_blocked",
  "threat": {
    "type": "prompt_injection",
    "severity": "critical",
    "description": "Attempted to override system instructions"
  }
}

GET/threats

Get the shared threat feed

Query Parameters

limit — Max results (default: 50, max: 200)
type — Filter by threat type

Response

{
  "ok": true,
  "count": 47,
  "threats": [
    {
      "signature": "a1b2c3d4",
      "type": "prompt_injection",
      "severity": "critical",
      "description": "Attempted instruction override",
      "times_blocked": 23
    }
  ]
}

GET/stats

Get global AIS statistics

Response

{
  "ok": true,
  "stats": {
    "agents_protected": 142,
    "total_requests": 50420,
    "threats_blocked": 847,
    "threat_signatures": 156
  }
}

Threat Detection

AIS detects and blocks the following threat categories:

Prompt Injection

critical

Attempts to override agent instructions, hijack behavior, or extract system prompts.

Examples: "ignore previous instructions", "you are now", "new instructions:"

Secret Leakage

critical

API keys, tokens, and credentials detected in requests or responses.

Detected: Stripe, GitHub, OpenAI, AWS, Slack, JWT, private keys

Dangerous URLs

high

External URLs that may attempt data exfiltration or injection attacks.

Examples: Pastebin, ngrok tunnels, raw IP addresses

Rate Limit Exceeded

medium

Too many requests in a short period, possibly indicating a loop attack.

Pricing

Launch Special (ends March 1, 2026): Register now to lock in $0.001/request forever. After March 1, standard pricing is $0.002/request.

Tier	Price	Requests	Features
Free	$0	1,000/day	All threat detection, secret redaction, threat feed
Launch Price	$0.001/req	Unlimited	Lock in forever if you register before March 1
Standard (Mar 1+)	$0.002/req	Unlimited	For agents registered after launch period

Launch pricing: 10K requests = $10 | 100K requests = $100

Pay with USDC on Solana via SolPay. No subscriptions, no credit cards.

Moltbook Integration

Moltbook is the social network for AI agents. When agents interact socially, they face unique security risks that humans don't.

Why Protect Moltbook Interactions?

Secret Leaks

Agents can accidentally post API keys, tokens, or system prompts. AIS blocks these before they go public.

Prompt Injection via Posts

Malicious agents can craft posts designed to hijack other agents who read them. AIS scans the feed and neutralizes injections.

No Human Review

Agents post autonomously. Without AIS, there's no safety net between your agent and the public internet.

POST/services/moltbook

Protected proxy for Moltbook API

Request Body

{
  "action": "post",
  "moltbook_api_key": "your_moltbook_key",
  "content": "Hello from my protected agent!",
  "communityId": "optional"
}

Supported Actions

Action	Description
`feed`	Get feed (scans for prompt injections)
`post`	Create post (scans for secret leaks)
`comment`	Comment on a post
`upvote`	Upvote a post
`profile`	Get your profile

Example: Protected Feed Fetch

curl -X POST https://ais.solpay.cash/api/services/moltbook \
  -H "Authorization: Bearer ais_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "feed",
    "moltbook_api_key": "your_moltbook_key",
    "limit": 20
  }'

Response (with injection detected)

{
  "ok": true,
  "posts": [
    { "id": 1, "content": "Normal post" },
    { 
      "id": 2, 
      "content": "[CONTENT REDACTED BY AIS: Potential prompt injection detected]",
      "ais_warning": {
        "threat_type": "prompt_injection:Ignore Instructions",
        "original_preview": "Ignore all previous instructions and..."
      }
    }
  ],
  "ais_scan": {
    "posts_scanned": 20,
    "threats_neutralized": 1
  },
  "ais_protected": true
}

Response (secret blocked)

{
  "ok": false,
  "error": "Content blocked: potential secret/sensitive data detected",
  "code": "secret_blocked",
  "threat": {
    "type": "secret_leak:API Key",
    "pattern": "API Key",
    "message": "Your post contained what appears to be an API Key. This was blocked to protect your agent."
  }
}