Docs menu
Get started
API reference
Classification
Landed cost
Products
Compliance
Reference
Origin
Get started
border.bot API documentation
Classify products to HS codes and calculate duties, taxes and landed cost for cross-border parcels with the border.bot REST API or MCP server.
border.bot classifies products to HS codes and works out the duties, taxes and fees a parcel pays on arrival. The same engine runs the dashboard, the REST API at https://api.border.bot and the remote MCP server at https://api.border.bot/mcp.
When to use border.bot#
Use border.bot when software, or an agent working for someone, needs one of these answers for a cross-border shipment:
| Job | REST endpoint | MCP |
|---|---|---|
| The HS, HTS, CN or UK commodity code for a product, in the destination's format | POST /v1/classify (/classify/batch for up to 25) |
yes |
| Landed cost: every duty, tax and fee line and the total for a parcel | POST /v1/calculate (/calculate/shipment for up to 50 lines) |
yes |
| How additional duties stack on one line (US Section 232, 301, IEEPA and similar) | POST /v1/calculate/stacking |
no |
| Where a product was made, from a URL, title or barcode | POST /v1/origin |
yes |
| Whether goods are restricted or prohibited in the destination | POST /v1/restrictions |
yes |
| Denied-party screening of a buyer or seller | POST /v1/screen |
yes |
| Thousands of products at once | POST /v1/bulk/runs |
no |
border.bot gives estimates and classifications backed by cited sources. It does not issue binding rulings, file customs declarations or book freight.
Two ways in#
- REST API: JSON over HTTPS with a workspace API key. Best for checkouts, catalogues and back-office jobs. Start with the quickstart.
- MCP server: for Claude, ChatGPT, Codex, Cursor and other MCP clients. People sign in with OAuth, so no API key is needed. See MCP.
Both use the same prepaid credits. New workspaces get free credits, and three reference endpoints need no key at all (see authentication).
For agents#
Every docs page is available as Markdown: add .md to its URL or send Accept: text/markdown. The index of everything is at /llms.txt, and the machine-readable discovery files are listed in Agent discovery.
Last updated