#33
Monthly Rank
#59
Lifetime Rank
Share
6 September 2026
Vyaas Docket

Vyaas Docket

AI docketing for Indian litigators: drop a file and it is docketed, ask the book anything, research 12 million judgments with citations.

Gallery

About the Project

Vyaas is a docketing suite for a litigation practice in India, built around three AI features that take the clerical work out of running a chamber. It runs from one folder on one machine, and nothing leaves the firm except the calls to the Claude API.

  1. Document Drop. Drag any PDF, Word or Excel file anywhere onto the screen. Vyaas reads it, works out which matter it belongs to, which folder it should sit in and what it should be called, and shows the proposal for a one-click confirm. Scanned PDFs with no text layer are transcribed page by page. Nothing is filed until the user approves it.

  2. New matter from documents. Drop the petition, the order, the FIR or the whole bundle for a new brief. Vyaas classifies the case type, fills the register form with the particulars that case type needs (parties, forum, case number, coram, stage, client side), drafts the first diary entry and files each PDF into the right folder of the new matter. Every field is editable before anything is saved.

  3. Ask Vyaas. A copilot over the live docket. It sees the whole book on every question, so it can answer "what is listed this week and which files need a brief", "which diary entries are overdue and whose are they", or "summarise the order sheet of my most urgent file". Answers keep streaming while the user works in other tabs, and several conversations can run at once. It never invents matters, dates or orders; it only reads the book as it stands.

  4. Research. A case-law chat over the Open India Law corpus (Vaquill AI, CC BY 4.0): 12 million judgments across the Supreme Court and 20 High Courts. Chosen courts are downloaded and indexed locally into SQLite full-text search.

Claude works like a junior would: it runs several targeted searches, reads the operative portions of the judgments it finds, and answers with citations.

Every cited case links to its source PDF. Research can be started from inside a matter, so the file's own documents are part of the context. Runs live on the server, so a tab switch or reload never loses an answer.

Underneath sits a complete docketing system: a matter register by case type with particulars, order sheet, document folders and tasks; a cause list grouped by court; a month diary; a limitation and tasks board; live Delhi High Court case-status lookup; clients, workload reports, team chat and user management. The AI features are useful only because that book exists, and the book stays current because the AI features make filing painless.

Model choice is a setting: a balanced tier on Claude Sonnet 5 for everyday use, or a best tier on Claude Opus 5 for research and intake. Usage and cost per feature are shown in Settings.

Practice Areas

Key Features

  1. Document Drop: drop any PDF, Word or Excel file and Claude files it into the right matter, folder and name, after a one-click confirm

  2. New matter from documents: the register form, first diary entry and folder filing are generated from the case PDFs, including scanned ones

  3. Ask Vyaas: a copilot that reads the live docket and answers questions about hearings, limitation, order sheets and housekeeping

  4. Research: case-law chat over 12M+ Indian judgments indexed locally, with search and read tools and a source PDF link for every citation

  5. Scanned-document reading: PDFs without a text layer are transcribed page by page on Claude Sonnet 5

  6. Model tiers and cost tracking: Sonnet 5 or Opus 5 per feature, with usage shown in Settings

  7. Full docketing underneath: matter register, cause list, calendar, limitation board, live Delhi High Court status, clients, reports, team chat, user management

  8. Local-first: one data folder, hashed sign-ins, nothing sent anywhere except the Claude API