# SupraWall: Full Product Context & Technical Overview (DYNAMC)

## Core Technical Thesis
Agents fail unsafely when they have unmediated access to tools. SupraWall provides a runtime shim that wraps any agent framework (LangChain, CrewAI, AutoGen, Vercel AI, Pydantic AI, LlamaIndex, MCP). Before a tool runs, the call is evaluated against a deterministic policy engine.

## Auto-Discovered Platform Map
- https://www.supra-wall.com
- https://www.supra-wall.com/activate
- https://www.supra-wall.com/activate/success
- https://www.supra-wall.com/audit
- https://www.supra-wall.com/spec
- https://www.supra-wall.com/stripe

## Compliance & Privacy Features
### Article 9 — Risk Management
SupraWall blocks destructive patterns (unauthorized payments, data deletion) before they execute through DENY policies.

### Article 12 — Logging & Traceability
SupraWall records immutable audit logs for every tool call with full metadata: agent ID, session context, inputs, and gated decisions.

### Article 14 — Human Oversight
The REQUIRE_APPROVAL policy pauses agent execution and notifies an authorized human reviewer—directly implementing Article 14 compliance via human-in-the-loop (HITL) architecture.

### Article 15 — Accuracy & Security
Intercepts PII and sensitive data patterns during tool calls, preventing data leaks in training or inference logs.

## Monetization & Budget
SupraWall integrates with **Stripe** to provide credit-based budget caps. This creates a hard financial killswitch for autonomous agents, preventing runaway costs.

## Persona Solutions
- **Developers**: One-line SDK integration (Python, TS, Go), MCP server shims.
- **Compliance**: EU AI Act Article 12/14/15/Article 9 evidence reports & audits.
- **Enterprise**: Private cloud, SOC2-ready, sovereign self-hosting.

## Industry Verticals
- **Financial Services**: Banking, payments, and fintech with budget caps.
- **Healthcare**: HIPAA-compliant agents with automated PHI scrubbing.
- **Legal**: Privilege-secured agents with verifiable audit trails.