Nveil is an open-source, no-code AI platform that lets users upload data and describe what they need in plain language, then generates reproducible, verifiable visualizations and analyses. Unlike general-purpose AI tools, Nveil uses a deterministic engine where an LLM only translates intent into a code pipeline executed by proven libraries like Pandas, NumPy, and Apache ECharts, ensuring mathematically correct results. It supports diverse file types including clinical data, time series, and images, making it suitable for environments where accuracy is critical.
Funding
Funding not disclosed
Founders
Product
Problem
General-purpose AI tools like ChatGPT and Claude generate plausible-looking but potentially incorrect results, fabricating numbers, dropping decimals, and confidently presenting averages they never actually computed. This makes them unreliable for data science tasks in professional settings such as board meetings, clinical trials, or financial closes, where mathematical accuracy is non-negotiable.
Solution
Nveil provides a trustable AI platform for data science that processes, analyzes, and visualizes data with real math, not guesses. Users upload their data and describe what they need in natural language; Nveil's LLM translates that intent into a deterministic data pipeline, which is then executed step by step in a sandbox. Every transformation, aggregation, and visualization is defined in code and produces the same output every time, ensuring reproducibility and verifiability. The platform orchestrates industry-standard tools like Kedro for pipeline orchestration, Apache ECharts for 2D visualization, VTK and Trame for 3D scientific rendering, and Pandas, NumPy, and Scikit-learn for data processing and statistical analysis. This approach guarantees that every number traces back to a transformation, making results auditable and defensible.
Target Audience
Primary users are engineers, researchers, and data scientists in university hospitals, financial institutions, and research labs who need accurate, auditable data analysis and visualization without writing code.
Features
- Deterministic engine: LLM only translates intent; all computations are executed by code, eliminating hallucinated numbers and ensuring mathematical correctness
- Reproducible pipelines: Run the same analysis with new data and get identical logic and results, with no drift
- Verifiable outputs: Every figure traces back to a specific transformation, with lineage tracking and audit trails
- Broad file format support: Ingests CSV, TSV, XLSX, XLS, ODS, JSON, PARQUET, DICOM, MHD, EDF, PNG, JPG, TIFF, BMP, WEBP, GIF, MAT, and XML files
- No-code interface: Users upload data and describe their needs in plain language, with no coding required
- Industry-standard stack: Built on Kedro, Apache ECharts, VTK, Trame, Pandas, NumPy, and Scikit-learn for production-grade reliability