Data & AI tools

For the work around a database or a model — schemas, notebooks, embeddings and training data.

These read your files, not your servers. A browser can only speak HTTP and WebSocket — it has no raw TCP sockets — so no web page can connect to MySQL or Postgres directly. Anything that claims otherwise is relaying your credentials through a server you don't own. These tools work on what you already have on disk: a schema dump, an .ipynb, an embedding export, a JSONL file. Nothing is uploaded.

SchemaViz

Turn MySQL, Postgres, SQLite or SQL Server DDL into an interactive ER diagram — drag, trace and export.

Open tool →

NotebookView

Open any Jupyter .ipynb with markdown, code, tables, tracebacks and plots intact — no kernel needed.

Open tool →

VectorLens

Project, cluster and search embeddings from any vector store — PCA plot plus exact cosine similarity.

Open tool →

DataSetInspect

Validate a fine-tuning JSONL before you pay to train on it — broken rows, duplicates, length outliers.

Open tool →