Step 1 of 4 · The mapTax authority › Data, IT & Analytics
The mapTax authorityData, IT & Analytics
Data, document AI & AI models
covered · 14 open projects
The data platform, analytics, document AI, language models, benchmarks and AI governance.
Start with
A benchmark to test a model on returns, and an embedding model tuned to French tax text.
Its 14 projects
| Project | Kind | Jurisdiction | Last update | Stars |
|---|---|---|---|---|
| TaxCalcBenchBenchmark that tests frontier AI models on US tax return calculations. | Benchmarks | US | Oct 2026 | 127 |
| Lemone-Embed LEmbedding model for search over French tax texts, fine-tuned from multilingual-e5-large. | Models | FR | — | — |
| docTRDeep-learning OCR library for detecting and recognising text in documents. | Automations | Global | Oct 2026 | — |
| ExtractThinkerDocument intelligence library that uses LLMs to extract structured data. | Automations | Global | Sep 2026 | — |
| Lemone mBERT taxmBERT model adapted to French tax language, for clustering and semantic search. | Models | FR | — | — |
| LLM-Tax-AbuseCode and US tax-strategy data from the paper on whether LLMs can identify and generate tax abuse.Licence unknown | Benchmarks | US | Oct 2025 | — |
| MinerUTurns complex PDFs and Office documents into structured, LLM-ready text.MinerU Open Source License | Automations | Global | Sep 2026 | — |
| Model Context ProtocolThe open standard for connecting AI agents to tools, data and sources. | MCP | Global | — | — |
| PaddleOCRMultilingual OCR and document-AI toolkit that turns PDFs and images into structured data. | Automations | Global | Sep 2026 | — |
| PolicyBenchBenchmark of how well frontier AI models estimate household tax and benefit outcomes without tools. | Benchmarks | Global | Oct 2026 | — |
| QWED-TaxDeterministic check, built on the Z3 solver, that verifies AI-generated tax decisions before they run. | APIs | US · IN | Sep 2026 | — |
| SARAStatutory reasoning benchmark built on the US Internal Revenue Code. | Benchmarks | US | — | — |
| SteuerLLM28-billion-parameter language model for German tax law analysis, from the SteuerLLM paper.CC BY-NC 4.0 | Models | DE | — | — |
| sydonia-toolkitSQL compiler and reference model for ASYCUDA World (SYDONIA) customs databases, read-only, with a sandbox. | Automations | Global | Aug 2026 | — |
Every project here is maintained and serves a tax authority. See the team's projects in the list.
On Learn
- BenchmarkGlossary term