ATLAS (AI Transport Logistics Agent Standard) is an open-source MCP server built for enterprise logistics. It runs inside your security perimeter and gives AI agents deep operational context — without your data ever leaving your infrastructure.
Connect ATLAS to your TMS, ERP, email, and documents. Any MCP-compatible AI agent (Claude, GPT, or custom) can then query your own data to negotiate rates, answer shipment questions, or run procurement workflows — from context that actually belongs to your company.
The Problem It Solves
Enterprise logistics companies have years of operational data: carrier relationships, rate history, shipment records, contracts. AI agents need this context to be useful. But sending raw operational data to external cloud services is a non-starter for legal and compliance teams.
ATLAS fixes this. Your data stays on your infrastructure. Agents get the context they need via the MCP protocol.
Key Features
• Zero data egress — all processing happens on your infrastructure • Logistics-native data models — Shipment, Carrier, Rate, Route, Document, Event • Connectors — Email (IMAP/Exchange), Google Drive/S3, REST API (TMS/ERP) • Standard MCP interface — works with Claude, GPT, and any MCP-compatible agent • Docker deploy — running in minutes with a single command • Open source — Apache 2.0, inspect every line
Use Cases
Carrier Negotiation — Agent queries your historical volumes with a carrier before negotiating rates. Negotiates from knowledge, not guesswork.
Shipment Tracking — "Where is shipment #12345?" — Agent pulls status from your TMS via ATLAS and answers instantly.
Procurement — "Best 3 carriers for reefer transport to Ukraine?" — Agent pulls your own performance data and makes a data-driven recommendation.
Server Config
{
"mcpServers": {
"atlas": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-p",
"3000:3000",
"-v",
"./config.yml:/app/config.yml",
"cargofy/atlas:latest"
]
}
}
}