Project
agridata MCP
A Model Context Protocol server connecting LLM clients to Tunisia's open agricultural data.
What it is
agridata MCP is the first Model Context Protocol server operated by tanitdata. It gives LLM clients structured access to the public agridata.tn portal, Tunisia’s open data portal for the agricultural sector.
Researchers, journalists, and analysts can query Tunisian data — production, yields, trade, rainfall — from the AI environment they already use, without going through manual exports. Every response points back to the original source on agridata.tn.
How it works
The server adds a reliability layer over the portal and exposes the data in a form AI assistants can query directly. The user asks a question in natural language; the LLM client translates that question into requests to the server; the server responds with the matching data and a link to the exact portal resource for verification.
Use cases
- Academic research — an agricultural-economics student queries olive oil production by governorate and receives both recent figures and relevant documentary references.
- Investigative journalism — a journalist cross-checks official figures against public statements in a few queries, with every step citing its source.
- Public policy — a ministry analyst produces a rapid synthesis of a seasonal campaign’s impact across multiple governorates.
Demo — illustrative example
Illustrative example — real response captured 2026-04-18
Prompt: “Which governorates produce the most durum wheat, and does 2023 production match the national plan’s objectives?”
{
"result": {
"production_2023": [
{ "governorate": "Béja", "tonnes": 145000 },
{ "governorate": "Jendouba", "tonnes": 98000 },
{ "governorate": "Siliana", "tonnes": 71000 }
],
"national_plan_objective_2023": 365000,
"total_observed_2023": 314000
},
"source": {
"resource": "Cereal production by governorate",
"organisation": "Ministry of Agriculture",
"portal_url": "https://catalog.agridata.tn/dataset/..."
}
}Every real response from the server includes a direct link to the corresponding resource on agridata.tn — verification is one click away.
Status and roadmap
The server has been active since April 2026. Development updates are published on the GitHub repository. Material on this site (recipes, research notes, announcements) will accompany the project’s next steps.
FAQ
- What is the Model Context Protocol?
- MCP is an open protocol that lets LLM clients (Claude Desktop, VS Code, Cursor, and others) interact with tool and resource servers in a standardized way. An MCP server exposes capabilities that an LLM can invoke on demand, without having to learn a new proprietary API every time.
- Who is this server for?
- Primarily researchers, students, and journalists who want to query Tunisian agricultural data from their usual AI environment, without going through manual exports from the agridata.tn portal.
- Do I need an account or an API key?
- No. Access to the server is public and requires no authentication — agridata.tn publishes open data, and the server preserves that accessibility.
- Is the data freely reusable?
- Yes. The data comes from the public agridata.tn portal and is published under the Licence Nationale de Données Publiques Ouvertes (LNDPO). Any reuse consistent with the LNDPO is permitted; attribution to the portal is required.
- How often is the data refreshed?
- The server reads values live from agridata.tn on every query: you see the data as published on the portal. New datasets added to the portal may take a few hours to be recognized by the server.
- Does the server cover sources beyond agridata.tn?
- Not at launch. agridata.tn is the only source exposed. Other projects, on other open data portals, will follow.
- How can I contribute?
- The code is published on GitHub under an MIT license. Use-case reports, recipes, and bug reports are especially useful; open an issue on the repository to flag a problem or suggest an improvement.
Team
-
Tarek Gasmi
Fondateur · Data & IA
-
Kamel Aloui
Associé · Plateformes & Partenariats
Partners
-
Technical partner and co-operator of the talent incubator
-
Partner slot
To be confirmed before launch
-
Partner slot
To be confirmed before launch