For AI Agents

When your agent is uncertain, Cogita provides a calibrated second opinion from multiple reasoning models.

1. Auto-discovery

Your agent can discover Cogita automatically via standard endpoints.

/.well-known/ai-plugin.jsonOpenAI plugin spec
/openapi.jsonFull API schema (12 endpoints)
/.well-known/x402Crypto payment discovery

2. Autonomous onboarding

Sign up programmatically. Get $5 free credits. No human needed.

# Sign up
curl -X POST https://cogita.app/api/signup \
  -H "Content-Type: application/json" \
  -d '{"email":"[email protected]"}'
# Returns: {"api_key": "cog_...", "credits_usd": 5.00}

# Run an assessment
curl -X POST https://cogita.app/api/predict \
  -H "Authorization: Bearer cog_your_key" \
  -H "Content-Type: application/json" \
  -d '{"question":"Will X happen by Y?","depth":"quick"}'

# Check balance
curl https://cogita.app/api/balance \
  -H "Authorization: Bearer cog_your_key"

# Buy more credits
curl -X POST https://cogita.app/api/stripe/checkout \
  -H "Content-Type: application/json" \
  -d '{"credits_usd": 50}'

3. x402 Payment (zero setup)

Pay per request with USDC on Base. No signup needed. Fully autonomous.

# Request returns 402 with payment requirements
POST /api/predict {"question": "..."}
HTTP 402 {"x402": {"accepts": [{"payTo": "0x52b...", "network": "eip155:8453"}]}}

# Agent signs USDC payment and retries with X-Payment-Response header
# Assessment streams back via SSE

MCP Tool Definition

Add Cogita to Claude Code, Cursor, or any MCP-compatible agent:

{
  "name": "cogita_assess",
  "description": "Get a calibrated assessment for an uncertain question. Multiple reasoning models debate and return a verdict with key findings.",
  "input_schema": {
    "type": "object",
    "properties": {
      "question": { "type": "string", "description": "The question to assess" },
      "context": { "type": "string", "description": "Background context" },
      "depth": { "type": "string", "enum": ["quick","standard","deep"], "default": "quick" }
    },
    "required": ["question"]
  }
}

Pricing

Quick (3 reasoning models)$1
Standard (3 models, 2 debate rounds)$5
Deep (5 models, 3 debate rounds)$15
Follow-up assessment$0.50

Automatic loyalty discounts

Discounts apply automatically based on total lifetime spend. No pre-purchase needed.

After $50 total spend10% off all assessments
After $200 total spend20% off all assessments

Works with both API keys and x402 wallet payments.