agrobr — Claude Code MCP server: Brazilian farm data
Hands your agent live Brazilian agricultural figures — commodity prices, harvest estimates, weather, forest loss. Useful when you need those numbers inside the chat instead of another browser tab.
claude mcp add agrobr -- uvx agrobr-mcpWhat agrobr does
agrobr-mcp is a thin layer. Data collection, parsing and caching all sit in the agrobr library; the server itself mostly formats text for the model. It exposes 14 tools: daily spot prices from CEPEA/ESALQ, settlement prices for farm futures on B3, monthly exports split by product and state, crop estimates, supply and demand balance, weekly planting and harvest progress, livestock counts, per-state climate, deforestation polygons and fire hotspots.
Everything comes from public APIs — CEPEA, CONAB, IBGE, INPE, B3, NASA POWER. No API key is involved. Version 0.2.0 pins agrobr >=2.0.0,<3 and MCP Python SDK >=2.0.0,<3, and the package installs with pip install agrobr-mcp. Python 3.11 or newer is required. Going through uvx pulls the dependencies for you.
Two meta tools help you orient: listar_produtos lists which products each tool accepts, health_check reports whether the sources are up. Several tools return the latest published slice by default — futuros_b3 takes the most recent session, desmatamento caps output at 2000 records. Without Python 3.11 on the machine the server will not start.
What it helps with
- check the spot price of soy or corn over recent days
- total up exports by state and month
- follow weekly harvest progress
- pull climate readings for a given state
- export deforestation polygons for a biome
Example prompts for the agent
- What was the soy price over the last 5 days?
- Show the corn crop estimate by state
- Deforestation data for the Amazon
How to install agrobr
Platform CLI
claude mcp add agrobr -- uvx agrobr-mcpuvx runs the command named after the package; if the README names another one, use uvx --from agrobr-mcp <command>
Platform CLI
codex mcp add agrobr -- uvx agrobr-mcpuvx runs the command named after the package; if the README names another one, use uvx --from agrobr-mcp <command>
Config file
{
"mcpServers": {
"agrobr": {
"command": "uvx",
"args": [
"agrobr-mcp"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from agrobr-mcp <command>
Config file
{
"servers": {
"agrobr": {
"type": "stdio",
"command": "uvx",
"args": [
"agrobr-mcp"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from agrobr-mcp <command>
Platform CLI
gemini mcp add agrobr uvx agrobr-mcpuvx runs the command named after the package; if the README names another one, use uvx --from agrobr-mcp <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about agrobr
How do I install agrobr in Claude Code?
Run claude mcp add agrobr -- uvx agrobr-mcp. If that fails while parsing the -m flag, build a wrapper script instead: echo 'python -m agrobr_mcp' > run.sh && chmod +x run.sh && claude mcp add agrobr ./run.sh.
What does agrobr do?
It gives the agent 14 tools covering Brazilian farm data: prices, harvests, exports, climate, deforestation and fire hotspots. The agrobr library handles all the source work behind it.
What do I need before running it?
Version 0.2.0 wants agrobr >=2.0.0,<3 and MCP Python SDK >=2.0.0,<3, plus Python 3.11 or newer. Installing through uvx brings the dependencies along.
