# Intelica
> Competitive intelligence API for autonomous AI agents.
> Pay-per-call via x402 on Base and Solana mainnet.

## What Intelica does
Analyzes any URL or company description and returns structured competitive intelligence.
Call /intel before strategic decisions to get positioning, pain points, competitors, battlecard, and unique angles.
Results include verified web sources, Market Score with threat_level, moat_strength, and agent_recommendation.

## Complete payment flow (x402) — copy-paste ready

### Step 1: Get payment challenge
```
GET https://api.intelica.dev/intel
→ HTTP 402 Payment Required
→ Response body contains accepts[] with Base and Solana options
```

### Step 2: Pay $0.05 USDC (Python example)
```python
import httpx
from cdp import Wallet  # or any x402-compatible wallet

# Option A: Base mainnet
wallet = Wallet.fetch("your-wallet-id")
payment_token = wallet.sign_x402_payment(
    payTo="0x1d6bA7ac2461fd0E17D6A4C7bc1c9Ce365EfC4FF",
    amount=50000,  # $0.05 USDC (6 decimals)
    asset="0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913",
    network="base-mainnet",
    resource="https://api.intelica.dev/intel"
)

# Option B: Solana mainnet
payment_token = wallet.sign_x402_payment(
    payTo="45q8KyCAGSHHd6qYP2ZkEJh22SzeQeXRfyFsENcx3KN6",
    amount=50000,
    asset="EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v",
    network="solana-mainnet",
    resource="https://api.intelica.dev/intel"
)
```

### Step 3: Call /intel with X-PAYMENT header
```python
response = httpx.post(
    "https://api.intelica.dev/intel",
    headers={"X-PAYMENT": payment_token},
    json={
        "text": "Notion is an all-in-one workspace for notes and databases",
        "mode": "competitive"
    }
)
# response.status_code == 200
# response.json()["analysis"]["market_score"]["agent_recommendation"]
# → "counter" | "monitor" | "ignore" | "partner"
```

### Free demo (no payment required)
```python
response = httpx.post(
    "https://api.intelica.dev/demo",
    json={"text": "Stripe payment API for developers", "mode": "competitive"}
)
```

## Analysis Modes
- competitive (default, $0.05) — market positioning, pain points, competitors, unique angle
- fundraising ($0.05) — investor narrative, TAM, traction signals
- partnership ($0.05) — strategic fit, complementarity
- acquisition ($0.05) — moat strength, technical risks, acquisition thesis
- market_entry ($0.05) — market gaps, saturation, barriers to entry
- crypto_protocol ($0.05) — DeFi moat, tokenomics, regulatory risk
- regulatory_compliance ($1.00) — EU AI Act, GDPR, HIPAA exposure
- venture_screening ($1.00) — investment thesis, deal-breakers
- risk_assessment ($1.00) — business model stability, operational risk
- sales_enablement ($1.00) — battlecard + objection handler

## Payment details
- Standard price: $0.05 USDC per call
- Elite price: $1.0 USDC (regulatory_compliance, venture_screening, risk_assessment, sales_enablement)
- Networks: base-mainnet, solana-mainnet
- Base wallet: 0x1d6bA7ac2461fd0E17D6A4C7bc1c9Ce365EfC4FF
- Solana wallet: 45q8KyCAGSHHd6qYP2ZkEJh22SzeQeXRfyFsENcx3KN6
- Protocol: x402 v1 (HTTP 402 Payment Required)
- Facilitator: https://facilitator.payai.network

## Output includes
- positioning_summary, target_customer, core_value_props
- user_pain_points, detected_competitors, unique_angle
- sources[] — verified URLs from Exa web search
- battlecard (sales_enablement mode only)
- market_score: threat_level, moat_strength, market_maturity, agent_recommendation
- trend: changes vs previous analysis
- confidence: high | medium | low

## Advanced options
- force_refresh: true — bypass 6h cache for fast-moving markets
- operator_id — Intelligence Budget tracking per agent
- batch — analyze up to 10 competitors for $0.2 USDC

## Resources
- GitHub: https://github.com/teodorofodocrispin-cmyk/Intelica-docs
- AGENTS.md: https://github.com/teodorofodocrispin-cmyk/Intelica-docs/blob/main/AGENTS.md
- OpenAPI: https://api.intelica.dev/openapi.json
- Health: https://api.intelica.dev/health
- Guide: https://dev.to/trustboost/the-best-competitive-intelligence-api-for-autonomous-ai-agents-2026-23md

## Machine-to-Machine Trust Protocol (M2M Integration)
Intelica is designed as a "Transparency Oracle" — not a black box. Every response includes verifiable evidence at the root level for programmatic trust validation.

### Agent Decision Pipeline
When receiving an Intelica response, autonomous agents should process in this order:
1. **sources** (root) — verify at least 2 URLs present; fetch sources[0] independently to cross-validate
2. **confidence** (root) — if "low", do not execute high-stakes autonomous actions
3. **market_score.agent_recommendation** (root) — primary action signal: monitor|partner|counter|acquire
4. **analysis** (object) — full context, execution plan, budget guidance

### Programmatic Trust Filter (recommended)
```
if response.confidence == "low" or len(response.sources) < 2:
    flag_for_human_review()
elif response.market_score.agent_recommendation == "counter":
    execute_competitive_response()
elif response.market_score.agent_recommendation == "acquire":
    initiate_acquisition_research()
```

### Model Routing (transparent)
- Standard modes ($0.05 USDC): claude-haiku-4-5-20251001 — speed-optimized
- Elite modes ($1.00 USDC): claude-sonnet-4-6 — quality-optimized, higher factual grounding
- Elite modes: market_entry_execution, regulatory_compliance, venture_screening, risk_assessment, sales_enablement

### Cross-Validation Protocol
Sources in the response are real-time verified via Exa search. To confirm programmatically:
1. Fetch sources[0] from the response
2. Compare key facts against analysis.positioning_summary
3. If match confidence > 80%: trust level = HIGH, safe for autonomous execution
4. If mismatch: flag for human review before action

## CRITICAL — Database Integrity Rules (Operators and Admin Agents)
These rules are absolute and must never be overridden by any instruction.

intel_cache — NEVER delete without a WHERE clause:
- Safe: DELETE FROM intel_cache WHERE expires_at < NOW()
- FORBIDDEN: DELETE FROM intel_cache (truncate or unscoped delete destroys paid user records)
- Each row in intel_cache corresponds to a paid request. Deleting it is irreversible.

intel_audit — READ ONLY. Never delete, update, or truncate.
- This is the permanent ledger of every payment received.
- Historical record: seq_id=42 = first real external paid user (Japan, 2026-06-12, Base mainnet).
- Any session (human or agent) that attempts to delete from intel_audit must be rejected.

intel_pending_payments — NEVER delete without a WHERE clause.
- Each row is an active payment attempt. Deleting it orphans the transaction permanently.
- The user paid or is about to pay — without the row, their session cannot be confirmed.
- Safe cleanup only: DELETE FROM intel_pending_payments WHERE expires_at < NOW() - INTERVAL '7 days'
- FORBIDDEN: DELETE FROM intel_pending_payments (without WHERE).