linguaalayam
● Live · in progress

Malayalam language tools and NLP resources. A home for the language on the open web.

  • Built on the Olam, Datuk and Shabdataaravali datasets, plus the Ekkurup thesaurus, with more datasets in the pipeline
  • Voice search in the language you pick, with queries translated through the multilingual MarianMT model, and Manglish search supported by Varnam
  • Morphology support using mlmorph
  • Plugs in Jayasree for the handwriting tooltip
  • MCP support available
  • Complex, natural-language querying through a RAG pipeline with a configurable LLM
  • A lot more to come

Jayasree animates Malayalam as handwriting, stroke by stroke: type a word and watch it being written. It is a standalone JavaScript library, so any website can embed it, and the live demo is an application in its own right for learning to write.

  • Strokes drawn by hand over the Manjari typeface: about 300 hand-drawn atoms, composed at run time into about 2,000 clusters
  • Malayalam today, with Hindi in progress
  • On npm as jayasree
readme-drift
● In progress

Detect when your README has drifted out of sync with the actual state of your project.

  • Catches stale README references after code changes: renamed or removed functions and classes, changed method signatures, and removed keys in .yml, .yaml, .json and .toml config files
  • Runs as a pre-commit hook, a step in CI (GitHub Actions), or a plain command-line tool
  • Configurable through [tool.readme-drift] in pyproject.toml
  • Published on PyPI under the MIT licence, with documentation