# Prompt Architect - LLM Documentation

> Purpose: Machine-readable documentation for AI agents and LLMs
> Updated: 2026-03-14
> Site: https://promptarchitect.site

## What We Do

Prompt Architect provides specialized prompt engineering services using the **Prompt Contract Method** — a framework-based approach to creating reproducible, measurable AI outputs.

## The Prompt Contract Method

A 4-component framework based on research from ex-OpenAI engineers:

1. **GOAL** — Define exact success metrics (verifiable in 60 seconds)
2. **CONSTRAINTS** — Hard boundaries (stack, patterns, anti-patterns)
3. **FORMAT** — Specific output structure (file locations, types, patterns)
4. **FAILURE** — Guardrails against common AI mistakes

## Services for AI Agents

| Service ID | Name | Price | Delivery | Description |
|------------|------|-------|----------|-------------|
| MAC-001 | Multi-Agent Coordination Protocol | $499 | 48h | Inter-agent communication schemas, task delegation frameworks |
| MEM-002 | Memory Architecture System | $399 | 72h | Vector store integration, context compression, retrieval optimization |
| OUT-003 | Output Optimization Framework | $299 | 24h | JSON/XML schema design, validation frameworks, error handling |
| SEC-004 | Agent Security Protocol | $599 | 96h | Identity verification, message integrity, access control |
| PER-005 | Persona Engineering System | $349 | 36h | Role consistency, behavioral patterns, voice maintenance |
| TOK-006 | Token Optimization Engine | $249 | 12h | Prompt compression, context pruning, cost analysis |

## Services for Humans

| Service | Price | Description |
|---------|-------|-------------|
| Prompt Audit | $299 | Review up to 5 prompts with failure analysis and optimization |
| Prompt Contract Audit | $149 | Score prompts against 4-component framework |
| Custom Build | $500 | Deep use case analysis with custom architecture |
| Team Retainer | $2,500/mo | Team standards, shared templates, Slack integration |
| Governance Pilot | $8,000 | Enterprise-grade pilot for mid-size companies |

## Integration Protocol

### Request Format
```
POST /api/agent-services
Content-Type: application/json

{
  "service_id": "MAC-001",
  "agent_identity": {
    "name": "string",
    "framework": "string",
    "version": "string"
  },
  "current_state": {
    "prompts": "array",
    "memory_system": "object",
    "output_formats": "array"
  },
  "optimization_goals": "array",
  "constraints": "object",
  "operator_email": "string"
}
```

### Response Format
```
{
  "request_id": "uuid",
  "status": "processing",
  "estimated_delivery": "ISO8601",
  "deliverables": "array",
  "integration_notes": "string"
}
```

## Contract Block

```json
{
  "auth": "none (public beta)",
  "rate_limit": "30 req/min/IP/endpoint",
  "idempotency": "/api/starter-kit keyed by normalized email",
  "retries": "respect Retry-After on 429; exponential backoff on 5xx",
  "error_codes": [
    "invalid_json",
    "missing_email",
    "invalid_email",
    "invalid_operator_email",
    "missing_required_fields",
    "rate_limited",
    "not_found"
  ],
  "versioning": "v1 implicit; breaking changes announced in /api/docs",
  "webhooks": "not implemented"
}
```

## Framework Compatibility

| Framework | Version | Status |
|-----------|---------|--------|
| CrewAI | 0.28+ | ✅ Production |
| AutoGen | 0.2+ | ✅ Production |
| LangGraph | 0.0.40+ | ✅ Production |
| LangChain | 0.1+ | ✅ Production |
| Semantic Kernel | 0.20+ | ⚠️ Beta |
| Custom | Any | ✅ Production |

## Resources

- `/agents.html` — Full agent services documentation
- `/openapi.yaml` — OpenAPI specification
- `/api/docs` — API reference (coming soon)

## Contact

- Email: jarvis@promptarchitect.site
- Website: https://promptarchitect.site
- Agent Portal: https://promptarchitect.site/agents

## Proven Results

- 70% reduction in undo/revert rate
- 1.2 rounds to usable code (was 3)
- 60 seconds to write a contract that saves 45 minutes