HisaabMate

HisaabMate is a software tool designed to automate the process of extracting and organizing time logs from informal Hinglish notes. The app runs entirely offline on a user's local machine, ensuring complete privacy and avoiding any per-request API costs.

The primary extraction engine utilizes the Google Gemma 3 (gemma3:1b) model, running locally via Ollama. This model is constrained by a JSON Schema, enabling it to accurately parse Hinglish text and convert phrases like "2 ghante" or "aadha ghanta" into structured, schema-validated time entries. The app also includes a deterministic Hinglish regex parser that can run in-process, even if the Ollama API is offline.

Users can input their work logs in the form of .txt or text-based .pdf files, which are then parsed and sanitized in-memory. The app enforces a 5MB size cap and a 30-page PDF limit. It also detects and handles scanned or image-only PDFs directly within the application without storing any data on the device's filesystem.

HisaabMate comes with a Streamlit UI styled with Space Grotesk, Plus Jakarta Sans, and JetBrains Mono typefaces. The interface includes features like an Audit Readiness bar, a one-click Markdown Invoice Memo, and the ability to paste structured data directly into client communication channels. The app has undergone extensive testing, with 21 automated tests covering schema validation, data integrity, and UI workflows.

Developers have praised HisaabMate for its ability to recover missed billable hours, its offline functionality, and the complete control it gives over client data. While some users would prefer a subscription-based model, the majority would be willing to pay a one-time fee for a local desktop tool that ensures client privacy and handles bilingual Hinglish input.