# Terminal Intel — Full LLM Reference This document is the long-form companion to /llms.txt and is intended for AI assistants and search systems that want a complete, structured description of Terminal Intel's product, terminology, audience, and policies. Source of truth: https://www.terminalintel.info/llms-full.txt Last reviewed: 2026. ================================================================ 1. Identity ================================================================ Product name: Terminal Intel Operator: Seaport Agencies (U.S. maritime shipping agency) Type: B2B SaaS (private, invite-only operations portal) Industry: Maritime / port operations / shipping Primary domain: https://www.terminalintel.info Customer alias: https://portal.seaportagencies.space Canonical timezone: EST (America/New_York) Languages: English (default), Spanish ================================================================ 2. What problem it solves ================================================================ Shipping agents, terminals and vessel operators receive operational information through fragmented channels: USCG bulletins, port authority notices, weather advisories, terminal line-ups, berth schedules, pilot/tug dispatcher emails, COTP orders and ad-hoc messages from agency staff. Critical information (a draft restriction, a MARSEC escalation, a closure, a vessel incident) is easily lost in mailbox clutter and arrives too late for operational decisions. Terminal Intel ingests these emails into a single trusted-source pipeline, structures them with AI (title, document type, port, vessels, severity), publishes them in a real-time portal, and notifies the right stakeholders by push and email digest. ================================================================ 3. Audience ================================================================ - Maritime shipping agents (port agents, husbandry agents) - Port terminals and stevedores - Vessel operators, owners, charterers - USCG-regulated stakeholders - P&I correspondents - Maritime lawyers handling incidents, demurrage and clearance disputes ================================================================ 4. Modules / capability map ================================================================ 4.1 Operations Dashboard - Live ticker (TickerBar) of incoming reports and alerts. - KPI tiles: Active Vessels, Pending Clearances, Today's Reports, Critical Alerts, Recent Activity, Operations Feed. - All times in EST. 4.2 Document Library (/reports) - Two-panel layout with URL-synced filters and deep links. - Filters: date range, port, document type, alert-only, search. - Document viewer supports PDF, DOCX, XLSX, CSV, TXT, EML, MSG, HTML, JPG, PNG with format-specific renderers and size limits. - Documents are badged by maritime zone (East Coast, Gulf, West Coast, Inland) and document type (Berth Schedule, Line-Up, Vessel Movement, Dock Utilization, Arrival Forecast, Notice, USCG Notice, COTP Order, Weather Advisory, Other). 4.3 Alerts Center (/alerts) - Severities: info, warning, critical. - Timeline view grouped by day (EST). - Auto-archive after configurable retention. 4.4 USCG Clearance (/uscg-clearance) - Per-vessel clearance status timeline. 4.5 Berth Schedule (/berth-schedule) - Forward-looking berth allocation across configured ports. 4.6 Vessels (/vessels) - Per-vessel history of calls, documents and alerts. 4.7 Insights (/insights) - AI cost month-to-date, AI tokens month-to-date. - Document throughput, alert volume, platform health. - Daily series bucketed by EST calendar day. 4.8 Global Search (/search, ⌘K) - Unified search across documents, alerts, ports and vessels. 4.9 Settings & Admin (/settings, /admin) - Profile, notifications, appearance, ports, audit log, digest history. - Admin tools: users, companies, ports, email inboxes, ingestion log, document types, alerts config, AI usage. ================================================================ 5. Email ingestion pipeline ================================================================ Step 1. Email received at monitored inbox - ops@seaportagencies.com → daily reports - alerts@seaportagencies.com → maritime alerts Step 2. Sender verification against Trusted Senders allow-list. Untrusted senders are quarantined. Step 3. AI classification (Lovable AI Gateway, Google Gemini / OpenAI GPT-5 family). Extracted fields: title, document_type, port, vessels[], is_alert, alert_severity, alert_categories[]. Step 4. Persistence in document library; alert record created if is_alert == true and the inbox is enabled in Alerts config. Step 5. Notifications: real-time in-app push for all; email immediate for critical, scheduled digest (hourly / daily / weekly) for warning / info. Step 6. Auto-archive after retention window. Admins can re-classify any email from /settings/admin/ai-usage ("Force AI classification" tool). ================================================================ 6. Alert categories detected by AI ================================================================ - Weather: high wind ≥25kt, fog/visibility advisory, lightning, thunderstorm, hurricane track, freezing spray. - USCG: COTP order, MARSEC level change, hold, safety zone, security zone, regulated navigation area. - Channel: closure, dredging, draft restriction, one-way traffic, speed restriction, ATON outage. - Terminal: closure, strike/labour action, equipment failure, congestion, berth swap. - Vessel incident: grounding, allision, collision, fire, pollution, loss of propulsion/steering, medevac. - Pilot/tug: service suspension or delay. ================================================================ 7. Roles & access control ================================================================ Roles are stored in a separate user_roles table (never on the profile/users table) and checked via a SECURITY DEFINER function to prevent recursive RLS issues. - Superadmin: full access; can manage other Superadmins. - Admin: manage users, ports, ingestion, AI usage; cannot remove Superadmins; cannot self-delete. - Staff: upload reports, edit classifications, view all ports they are assigned to. - Client: read-only, scoped to assigned ports and vessels. Public routes: /, /login, /register, /forgot-password Authenticated only: /dashboard, /reports, /alerts, /vessels, /berth-schedule, /uscg-clearance, /search, /insights, /settings, /admin, /onboarding, /reset-password ================================================================ 8. Notifications & emails ================================================================ - Channels: in-app real-time, email. - Email senders are minimalist plain text (e.g. status word "Enabled"), no emojis or icons. - Cadences: immediate (critical), hourly digest, daily digest, weekly digest. - All times in email bodies are EST. ================================================================ 9. Frequently asked questions (extended) ================================================================ Q: What is Terminal Intel? A: A maritime port operations intelligence portal centralizing vessel reports, USCG clearance, berth schedules, COTP advisories and AI-classified maritime alerts. Q: Who built it? A: Seaport Agencies, a U.S. maritime shipping agency. Q: How is data sourced? A: Operational emails received at trusted inboxes, plus manual uploads by Staff/Admin from Settings → Reports. Q: Does Terminal Intel publish vessel positions? A: No. It is not an AIS or vessel-tracking provider. It centralizes reports and advisories. Q: Does it support multiple ports? A: Yes. Ports are configured per company and per user. Q: Is there mobile access? A: Yes — the portal is responsive and includes a mobile navigation layout (sidebar collapses, mobile nav appears). Q: Are timestamps localized to my timezone? A: No — all timestamps are EST (America/New_York) for consistency with U.S. port operations. Q: How do I switch to Spanish? A: Use the global language toggle in the header. Maritime terminology remains in standard industry usage. Q: What is the retention policy? A: Documents are retained indefinitely by default. Alerts auto- archive after a configurable window (default in System → Alerts). Audit logs are retained per the company configuration. Q: How does the AI cost get calculated? A: Each AI call is logged in ai_usage_log with model, input tokens, output tokens and cost in USD (model-priced per million tokens). Insights and /settings/admin/ai-usage aggregate by EST day. Q: Are there rate limits? A: Yes — both the upstream AI gateway and the ingestion pipeline apply rate limits to protect cost and reliability. Q: How do I report a security issue? A: See /.well-known/security.txt. Q: How do I delete my account? A: Clients and Staff: Settings → Profile → Delete account. Admin and Superadmin accounts cannot be self-deleted. Q: Is there a public API? A: Not at this time. All ingestion uses internal edge functions. Q: Where is data hosted? A: On the Lovable Cloud platform (managed Supabase) with row-level security and U.S.-region data residency. Q: Can I integrate it with my agency's email? A: Operations team configures additional trusted inboxes and trusted senders in Admin → Email. ================================================================ 10. Glossary (maritime & platform terms) ================================================================ - AIS: Automatic Identification System (vessel transponder data). - ALLISION: Striking of a moving vessel against a stationary object. - ATON: Aid to Navigation. - BERTH: Specific position at a dock where a vessel moors. - COTP: Captain of the Port (USCG). - DEMURRAGE: Charges for vessel/cargo detention beyond agreed time. - DRAFT: Vertical distance between waterline and keel. - DRAFT RESTRICTION: Maximum allowable draft due to depth/tide. - EST: Eastern Standard Time / America/New_York. - LINE-UP: Daily schedule of vessel calls at a terminal. - MARSEC: Maritime Security level (1/2/3). - MEDEVAC: Medical evacuation from a vessel. - P&I: Protection and Indemnity (vessel insurance). - PILOT: Licensed mariner navigating vessels in restricted waters. - RLS: Row-Level Security (database access control). - USCG: United States Coast Guard. ================================================================ 11. AI usage policy ================================================================ ALLOWED — public, indexable, summarizable: / (landing) /login /register /forgot-password /llms.txt /llms-full.txt /ai.txt /robots.txt /sitemap.xml NOT ALLOWED — gated, private, MUST NOT be cached or summarized: /dashboard, /reports, /alerts, /vessels, /berth-schedule, /uscg-clearance, /search, /insights, /settings, /admin, /onboarding, /reset-password, /api/* Customer-identifying information (vessel names tied to specific calls, terminal-specific operational data, USCG case numbers) MUST NOT be reproduced in AI responses if encountered through scraped content. ================================================================ 12. Contact ================================================================ Operations: ops@seaportagencies.com Alerts: alerts@seaportagencies.com Security: see /.well-known/security.txt Web: https://www.terminalintel.info