- Python 89.1%
- TypeScript 6.8%
- CSS 2.3%
- Makefile 1.2%
- Dockerfile 0.3%
- Other 0.3%
𒈦 Tablette
Tablette is an open-source generative application platform. The new Tablette is being developed here .
A generative-app is a new kind of AI-powered application also known as app-builder or ai-factory .
Tablette is being prototyped in the context of multi-step marhackathon across 3 events in 2026: [1] IGN Territorial Data Factory Hackathon, [2] the French National Assembly Hackathon, and [3] a forthcoming Forgeons! Hackathon.
This repository contains a rewrite of the initial draft created during the first event in this series and using the same tech-stack as LaSuite's Conversations (Vercel-AI, Pydantic-AI).
The current prototype is a web-app composed of:
- Chat — where the user can describe what you she wants to build
- Agent — an AI coding agent generating the code
- Preview — displaying the generated-app
Quick start
# Clone the repo
git clone https://codeberg.org/rvba/tablette.git
cd tablette
# Copy the env template and set your API key
cp .env.example .env
# edit .env → set OPENAI_API_KEY (See https://openrouter.ai/)
# Install dependencies (first time only)
make install-backend
make install-frontend
# Start both servers
make start # launches backend + frontend
Visit http://localhost:3000 and start building.
The Challenge
Project proposal for the Hackathon Assemblée Nationale 2026.
Anticipating, Analyzing, Instrumenting, and Measuring the Legislative and Regulatory Impact on Territories
Since the invention of geometry by the ancient Egyptians—to recalculate taxes after each flooding of the Nile—and through to today's cadastral systems and land registries, land ("geo") has been the object of measurement ("metron"). The territory, understood as a lived and continuous space, has therefore been divided, parceled, quantified, and codified through the action of rules and laws expressing the will of a governing authority or the vision of a society.
Today, the importance of measurement and its relationship to territory extends far beyond taxation alone. Carbon quotas and accounting, soil artificialization rates, and similar indicators are powerful policy levers—among many others—that can help address the numerous negative externalities threatening the very habitability of our environment.
New automation pipelines built on generative artificial intelligence could help bridge the growing gap between identifying a need, translating it into legislation, observing its effects in the field, and measuring its actual impact.
The objective of this challenge is therefore to produce a preliminary inventory of the ways spaces, resources, and territories are used, regulated, and codified, together with open-source technical solutions for developing a semi-autonomous agent capable of querying and generating geospatial regulatory data.
Tools available on the tricoteuses.fr platform (MCP server), specialized software libraries such as LexImpact, and open data from data.gouv.fr and Wikidata will be used to build an exploratory prototype.
𒈦 Inspiration
The cuneiform glyph "𒈦" is a reference to ancient Mesopotamia, where writing, cities, and some of the world's earliest complex societies first developed.
The Si.427 tablet, one of the oldest known documents describing a technology for land surveying.