Program
How can public data become usable, trustworthy, and AI-ready?
Open-data access, MCP tools, AI-ready public datasets, data audits, and source-grounded access infrastructure.
Public data is increasingly read through AI assistants, yet most open-data portals were never built for that interface — formats, metadata, and access patterns assume a human downloading a file. This program builds the connective layer that makes public datasets directly addressable by AI clients, and audits how ready those sources actually are. The aim is access that is immediate, verifiable, and grounded in the original source rather than a model's recollection.
Public Data
A Model Context Protocol server connecting LLM clients to Tunisia's open agricultural data.
Public Data
Ask Tunisia's national statistics in plain language — an MCP server over the INS data portal.
Public Data
A qualifier-aware MCP server over Tunisia's national energy statistics (2010–2026), derived from public ONEM reports.
Public Data
The first agentic map for Tunisian agricultural data — ask in natural language, get a chart, map, table, or short text in return.
AI & Governance
A political-economy analysis of LLM token flow: dependence is concentrating in the flow layer — the gates every inference request must clear — not in who produces models.
Researchers, institutions, and funders who want to contribute to this program — through data, a partner challenge, or a collaboration — can reach the studio via the Get Involved page.
