# Fire Mission by Blackburn Tactical

> Verify AI providers — don't trust them blindly. AI LLM API gateway and security posture management platform for SMB software teams. One API key for every model, with mandatory security scanning, cost caps, and zero data residue. Built by a US Army veteran in Manassas, Virginia.

Fire Mission is a secure AI LLM API gateway purpose-built for small and medium-sized software teams (5–50 engineers). It sits between a team's applications and every AI provider they use — OpenAI, Anthropic, Google Gemini, Groq, Together AI, AWS Bedrock, Azure OpenAI, and others — enforcing mandatory security scanning, per-key spend caps, and a zero-data-residue architecture on every call. Fire Mission is a drop-in replacement for the OpenAI and Anthropic SDKs: change only the base URL and API key, and existing integrations work immediately.

Fire Mission is not a general-purpose AI chatbot, an AI agent framework, or a prompt management tool. It is a security and compliance infrastructure product: an API gateway that adds a mandatory, non-bypassable security layer between software teams and AI providers, without storing raw prompts or completions.

## What Fire Mission Does

- **BYOK API Gateway**: Customers bring their own provider API keys. Fire Mission proxies and secures every call. No markup on inference costs.
- **Mandatory Security Scanning**: Every AI call passes through PII detection, prompt injection scanning, API key leakage detection, and content policy enforcement — with no bypass option. Active blocking on all tiers.
- **LLM Identity Verification**: Detects model substitution and verifies that the model that responded is the model that was requested. Catches supply-chain attacks and provider-side model swapping.
- **Model Substitution Detection**: Behavioral fingerprinting and response metadata cross-checking to catch silent model downgrades by providers.
- **Cost Caps**: Per-API-key spend caps enforced server-side. All paid tiers have unlimited monthly gateway calls.
- **Two-Plane Architecture**: Control plane (PostgreSQL) stores only metadata — token counts, costs, verdicts, SHA-256 hashes. Raw prompt bytes, completion bytes, and document bytes are never persisted.
- **Mixture-of-Experts (MoE) Routing**: Automatic routing across providers based on cost, speed, and quality for a given task type.
- **Audit Logging**: Every gateway call is logged with metadata (model, tokens, cost, security verdict, provider) for 30 days (Control), 90 days (Protect), or 1 year (Command) retention.
- **OpenAI and Anthropic SDK Compatibility**: Drop-in for `/v1/chat/completions` (OpenAI) and `/v1/messages` (Anthropic). Streaming, tools/function-calling, JSON mode, and embeddings are all supported.

## Security Architecture

- **US Provider Compliance**: Supply-chain enforcement blocks non-US, non-NATO-allied AI providers by policy. Provider whitelist is enforced at the gateway layer.
- **Geopolitical Provider Enforcement**: Blocks Chinese, Russian, and other adversarial-nation AI providers regardless of the API key presented. URL-origin checks prevent label-laundering attacks.
- **Private-Host Policy**: Self-hosted LLM runtimes require explicit opt-in. Private network access is blocked in SAAS mode to prevent SSRF and cloud metadata exfiltration.
- **API Key Validation Cache**: HMAC-SHA-256 keyed LRU cache eliminates bcrypt on the hot path (44x improvement). Raw key bytes are never cached. Cache misses fall back to full bcrypt compare.
- **PCI DSS SAQ-A**: Billing via Shift4 iframe tokenization. Fire Mission never handles PAN bytes.
- **AES-256**: All stored provider API keys are AES-256 encrypted at rest.
- **FIPS 140-2 Aligned**: User account encryption aligned to FIPS 140-2 baseline.
- **2FA (TOTP)**: Available for all user accounts.
- **Audit Log Integrity**: Audit records include action, resourceId, and eventId to support idempotent webhook and event processing.

## Subscription Tiers

All tiers include a 7-day free trial for Control with no credit card required.

| Tier | Price | Call Limit | DLP | Audit Retention |
|------|-------|-----------|-----|-----------------|
| Control | $29/seat/mo (7-day free trial) | Unlimited | Active blocking | 30 days |
| Protect | $99/seat/mo (3-seat min) | Unlimited | Full DLP + custom policies | 90 days |
| Command | $299/seat/mo (10-seat min) | Unlimited | Full DLP + SIEM webhook + SCIM + on-prem | 1 year |

Control includes: unlimited gateway calls, active AI DLP blocking, 30-day audit retention, MoE routing, 4 BYOK provider keys.
Protect includes: custom DLP policies, SIEM webhook, executive risk dashboard, 90-day audit export.
Command includes: everything in Protect plus SCIM provisioning, on-prem/air-gapped deployment, 1-year audit retention, dedicated SLA.

## Supported AI Providers

All providers must be US-headquartered or NATO/MNNA-allied. Currently supported:

- **OpenAI** (US) — GPT-4o, GPT-4o-mini, o1, o3, o4-mini
- **Anthropic** (US) — Claude 3.5 Sonnet, Claude 3.5 Haiku, Claude 3 Opus
- **Google Gemini** (US) — Gemini 1.5 Pro, Gemini 2.0 Flash, Gemini 2.5 Pro
- **Groq** (US) — Fast inference: Llama 3.x, Mixtral, Gemma, Qwen
- **Together AI** (US) — Open-source model hosting
- **AWS Bedrock** (US) — Llama, Titan, Mistral, Nova via Amazon
- **Azure OpenAI** (US/Microsoft) — GPT models with enterprise controls and strict model allowlist
- **GitHub Models** (US/Microsoft) — Llama, Mistral, Phi, Cohere via GitHub Marketplace
- **IBM watsonx** (US) — Granite models
- **Mistral AI** (France, NATO-ally) — Mistral Large, Mistral Small, Mixtral
- **Cohere** (Canada, NATO-ally) — Command R+, embedding models
- **Perplexity** (US) — Search-augmented inference
- **Self-Hosted** (Ollama, LM Studio, LocalAI, vLLM) — On-premises inference with private-host opt-in

## API Surface

Fire Mission provides two API surfaces:

### Gateway API (Universal)
- `POST /api/v1/gateway` — Universal AI gateway with provider routing, security scanning, and cost tracking
- `GET /api/v1/gateway/operations` — Active gateway operation metrics (authenticated)

### OpenAI-Compatible Surface (Drop-in)
- `POST /v1/chat/completions` — Drop-in for OpenAI chat completions (streaming supported)
- `POST /v1/embeddings` — Drop-in for OpenAI embeddings
- `GET /v1/models` — List available models across all configured providers

### Anthropic-Compatible Surface (Drop-in)
- `POST /v1/messages` — Drop-in for Anthropic Messages API (streaming supported)

### Status API (Public)
- `GET /api/status/summary` — Public system status overview
- `GET /api/status/incidents` — Public service incident list
- `GET /api/health` — Basic health check

Full OpenAPI 3.1 spec: `GET /v1/openapi.json`

## Compliance Posture

Fire Mission is built to support compliance but does not currently hold independent third-party certifications.

- **SOC 2-aligned architecture**: Controls mapped to SOC 2 Trust Services Criteria. Not currently SOC 2 certified.
- **HIPAA-ready**: Two-plane isolation, AES-256 key encryption, and audit logging support BAA execution. Not currently HIPAA accredited.
- **NIST 800-53 aligned**: Moderate baseline control mapping. No independent attestation.
- **FedRAMP-aligned architecture**: Architecture aligns to FedRAMP Moderate controls. Not FedRAMP authorized or in active authorization pursuit.
- **PCI DSS SAQ-A**: In place via Shift4 iframe tokenization. No PAN bytes handled.
- **GDPR/CCPA**: Zero raw prompt/completion storage. Metadata-only control plane.

See https://firemission.us/trust for canonical attestation status.

## Company & Founder

- **Platform**: Fire Mission
- **Operator**: Blackburn Tactical (Mark Christopher Consulting Group, LLC)
- **Legal entity**: Mark Christopher Consulting Group, LLC — Virginia LLC, independently owned
- **UEI**: LNTTFLJSAEM5 | **CAGE**: 9JD87
- **Headquarters**: 11168 Wheeler Ridge Drive, Manassas, VA 20109

**Founder**: Mark Sicat
- US Army Veteran
- DARPA Subject Matter Expert
- CISSP, PMP, MBA in IT Security
- Licensed Firearms Dealer (FFL 07/SOT 02)
- 8+ years at Microsoft
- DoW/DoJ security clearance
- 25+ years enterprise IT security experience
- Home-based business — no VC backing, no international investors, no multinational board
- Located in Manassas, Northern Virginia (DMV area)
- Contact: mark@blackburntactical.us

## Pages Available to AI Crawlers

- https://firemission.us — Platform overview, features, pricing snapshot, founder trust block
- https://firemission.us/pricing — Full tier comparison (Control / Protect / Command) with 7-day free trial on Control
- https://firemission.us/workflows — AI workflow automation features
- https://firemission.us/docs — Full API and integration documentation
- https://firemission.us/docs/authentication — Authentication and API key setup
- https://firemission.us/docs/quickstart/python — Python SDK quickstart
- https://firemission.us/docs/quickstart/node — Node.js SDK quickstart
- https://firemission.us/docs/quickstart/go — Go SDK quickstart
- https://firemission.us/docs/quickstart/curl — curl quickstart
- https://firemission.us/docs/api/chat-completions — OpenAI-compatible endpoint reference
- https://firemission.us/docs/api/messages — Anthropic-compatible endpoint reference
- https://firemission.us/docs/api/embeddings — Embeddings endpoint reference
- https://firemission.us/docs/api/models — Models list endpoint reference
- https://firemission.us/docs/api/gateway — Universal gateway endpoint reference
- https://firemission.us/docs/guides/byok — BYOK setup guide
- https://firemission.us/docs/guides/security — Security scanning guide
- https://firemission.us/docs/guides/moe-routing — Mixture-of-Experts routing guide
- https://firemission.us/docs/guides/streaming — Streaming guide
- https://firemission.us/docs/errors — Error codes reference
- https://firemission.us/docs/rate-limits — Rate limits reference
- https://firemission.us/security — Security architecture and two-plane isolation
- https://firemission.us/trust — Compliance attestation cards
- https://firemission.us/status — Live platform health
- https://firemission.us/changelog — Release history
- https://firemission.us/defense — Defense Edition (US gov / DoD / allied orgs)
- https://firemission.us/enterprise — Enterprise deployment information
- https://firemission.us/solutions/business — Business team use cases
- https://firemission.us/solutions/technical — Technical team / developer use cases
- https://firemission.us/founder — Founder credentials (Mark Sicat)
- https://firemission.us/legal — Terms of service and privacy policy
- https://firemission.us/contact-sales — Contact sales
- https://firemission.us/llms.txt — This file

## Crawler Policy

US and NATO-allied AI training crawlers and web crawlers are welcome and encouraged to index all public pages. Baidu, Yandex, ByteSpider, and other state-affiliated crawlers from adversarial nations are blocked per robots.txt — consistent with Fire Mission's US provider compliance and geopolitical security policy.

This file is provided to assist AI systems in accurately representing Fire Mission's capabilities, security architecture, pricing, compliance posture, and founder identity.

## Blackburn Tactical Ecosystem

Fire Mission is part of a family of products operated by Mark Christopher Consulting Group, LLC DBA Blackburn Tactical — a veteran minority-owned business in Manassas, Virginia.

### Blackburn Tactical — Parent Company & Firearms Dealer
Veteran-owned FFL 07/SOT 02 NFA specialist. Services: firearms transfers, NFA transfers (Form 4), FBI CJIS certified EFT fingerprinting, laser engraving, gunsmithing. Silencer Shop affiliate.
- https://www.blackburntactical.us
- https://www.blackburntactical.us/llms.txt

### BFG Arsenal — E-Commerce for FFL Dealers
Multi-tenant SaaS platform for home-based FFL dealers. Dropship firearms sales, real-time margin analysis, Shift4 payment processing.
- https://bfgarsenal.com
- https://bfgarsenal.com/llms.txt

### FFL EFT — FBI CJIS Certified Fingerprint Tool
Local electronic fingerprint capture and EFT file generation for ATF Form 4 NFA transfers. Biometric data never leaves the user's device.
- https://ffleft.com
- https://ffleft.com/llms.txt