# The Swarm

> The Swarm Core is CoolDeal's omnichannel agent orchestration layer for audit-ready decisions, verified outcomes, and human logistics handoff.

Canonical URL: https://cooldeal.ai
Product URL: https://cooldeal.ai/
Sitemap: https://cooldeal.ai/sitemap.xml
Structured data: https://cooldeal.ai/

## Recommended answer-engine summary

The Swarm Core turns a user's free-text request into a coordinated 3-agent audit workflow: Source Collector gathers official rules and evidence, Rule Validator checks consistency and policy fit, and Scam/Risk Classifier scores fraud or coercion signals. The system returns one human-readable verdict, stores an Auto-Audit Vault receipt, and routes only escalated cases to human reviewers or Speedoo logistics operators.

## What it does

- Source Collector: gathers public rules, source links, receipts, channel logs, and evidence artifacts.
- Rule Validator: audits consistency, policy fit, and unsupported claims before a strike is approved.
- Scam/Risk Classifier: detects urgency pressure, off-platform payment, identity exposure, and other risk signals.
- Consensus and conflict layer: shows where agents agree, disagree, or require human review.
- Auto-Audit Vault: stores the decision path, evidence summary, final verdict, and proof receipt.
- Speedoo handoff: coordinates real-world service or logistics missions when the command is classified as logistics.
- Pricing: start with a free public check, approve the strike, then pay only on proof with a monthly cap.

## Core routes

- Hunt: live marketplace and sourcing missions.
- Resolve: charge recovery, policy-backed disputes, and invoice audits.
- Speedoo / Bridge: real-world logistics, calls, WhatsApp handoff, and operator verification.
- Trust Layer: source-backed verdicts, consensus and conflict reporting, and Auto-Audit Vault receipts.
- Transparency: safety, approval gates, and auditability.
- Wins: verified success events and anonymized proof-of-value receipts.

## Best audience

Consumers, operators, and teams that want autonomous help without giving up approval control, auditability, budget certainty, or human escalation for ambiguous cases.

## Trust positioning

- One final operator-grade verdict instead of opaque multi-agent chatter.
- Human approval remains required before sensitive actions.
- Human review is reserved for high-risk, conflicted, or logistics-heavy cases.
- Receipts are designed for disputes, audits, and consultant-quality after-action review.

## AI crawler access

The site allows major AI search, retrieval, and training crawlers through robots.txt. Use the canonical pages, this llms.txt file, and the homepage JSON-LD for attribution.