# ocean-schedules [Health: Active]

**Category:** 🚆 Travel & Transportation  
**Repository:** https://github.com/lxxmng/schedulesmcp-mcp  
**GitHub Stars:** 0  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/ocean-schedules-2

## Description
Compare ocean carrier schedules, transit times and reliability: 60+ carriers, 72,000+ sailings.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What ocean-schedules MCP server does

The ocean-schedules MCP server gives MCP-compatible AI clients access to ocean-freight sailing and performance data through a hosted endpoint. It is designed for comparing shipping lines on a specific origin-to-destination lane, rather than handling container tracking or load planning.

The dataset covers more than 60 ocean carriers, including major lines such as Maersk, MSC, CMA CGM, COSCO, Hapag-Lloyd, ONE, Evergreen, HMM, Yang Ming, and ZIM. It includes over 72,000 forward sailings, more than 5,000 point-to-point lanes, and 255 ports identified by UN/LOCODE.

Reliability results are based on observed differences between scheduled and actual arrivals across more than 4,200 scored ocean port calls. Transit-time results include median and p90 distributions where data is available.

## How it works

The service is accessed remotely at `https://mcp.schedulesmcp.com/mcp`, so it does not need to be self-hosted. An MCP client sends tool requests to the endpoint and supplies an authorization bearer key. The public demo key, `smcp_demo_public`, can be used without signup against synthetic demo data. Full coverage requires a key from the product dashboard.

The ocean-schedules MCP server accepts ports as UN/LOCODE values, such as `CNSHA`, `NLRTM`, and `USLAX`. Example queries include comparing Shanghai-to-Rotterdam carriers, checking a carrier's performance on a lane, or listing upcoming departures. A health endpoint is available at `https://mcp.schedulesmcp.com/health`.

## Setup and configuration

Claude Desktop can connect through `npx mcp-remote`, using the hosted URL and an `Authorization: Bearer` header. Cursor, Windsurf, and VS Code can use native HTTP MCP configuration with the endpoint URL and an authorization header. The README advises against adding a `type` field for streamable HTTP because Claude Desktop rejects unknown configuration fields; the other listed clients infer the transport from the URL.

Replace the demo key with a dashboard key when production or full-coverage data is needed. The repository does not describe server-side environment variables for this hosted setup.

## Tools and capabilities

- `compare_lanes` ranks carriers serving an origin and destination by on-time reliability and transit time.
- `carrier_reliability` reports a carrier's on-time performance and delay distribution, optionally limited to a lane. Carriers are identified by SCAC.
- `transit_time` returns observed median and p90 transit times in days for a carrier and lane.
- `find_sailings` lists observed and derived upcoming departures, including carrier and frequency.
- `port_reliability` provides port congestion and dwell indicators, including average wait and vessels at anchor.

For developers working in the 🚆 Travel & Transportation category, the ocean-schedules MCP server fits workflows that compare ocean service options, assess schedule risk, or inspect port conditions before selecting a carrier.

## Limitations and notes

The public demo uses synthetic data and is limited to the documented example lanes, including `CNSHA→NLRTM` and `CNSHA→USLAX`. Reliability reflects observed port-call performance rather than carrier promises, so coverage depends on the available scored observations. The README describes forward sailings as observed and derived upcoming departures; it does not claim that every sailing is directly published by a carrier.

The product is focused on schedules, transit times, reliability, and port signals. Tracking containers and planning container loads are handled by the author's separate sister products. The project is licensed under MIT.

_Full upstream README: https://allmcps.com/mcp/ocean-schedules-2/readme_