Indian law.Actually answered.
Built for advocates who |
Research, draft, and verify in one place, with citations you can defend. Aware of BNS, BNSS, and BSA. Intake in your client's language.
What it looks like
The workspace, while you read this page.
A sandboxed slice of the Sanhita app, ten panes, looping live below. Watch the Assistant ground an answer in real Indian case law, then jump to Workflows, Drafter, Editor, Vault, and the rest. Click the sidebar to skip ahead.
Closed beta · India
What can I research for you today?
Bail · NDPS §37
Grounds for bail under Section 37 NDPS Act, recent HC trends?
Section 138 NI Act
Standard of proof for cheque dishonour under §138.
Writ · Art. 226
When can a High Court issue certiorari under Art. 226?
* For illustrative purposes only. The real workspace is materially deeper, India’s largest legal corpus, real grounding, real drafting. Join the waitlist to get the full experience →
What it does
Six surfaces. One workspace. Cited, or held back.
Research without dead-ends
Ask in your own words. Sanhita pulls the precedents that matter, ranked by relevance to the actual question, with a source rail you can scroll line by line.
Drafting that holds up
Cheque-bounce notices, plaints, replies, NDAs, lease deeds, employment letters. Generated from your matter brief, exported to Word with track changes intact.
Citations you can defend
Every authority that appears in an answer is checked against the live index before it renders. Multiple validator gates. If a case fails verification, the answer is held back rather than guessed.
Workflows that compound
Build a recipe once, run it across matters. Agreement audit, chronology, compliance gap, due diligence, focused summarizer. The same recipe runs again next month.
Speak your language
Hindi, Tamil, Bengali, Marathi, and more. Speak the matter the way your client described it. Sanhita transcribes, structures, and routes to the right surface.
Aware of the new codes
When you cite IPC, Sanhita shows BNS alongside. CrPC shows BNSS. IEA shows BSA. Two numbers in the chip, so a filing reads correctly under either reference.
How it works
Four steps to a citable memo.
You ask.
Type or speak the matter in plain language. No keyword juggling, no operator syntax. Sanhita reads intent the way an advocate phrases a question.
Sanhita retrieves.
Retrieval fans across the relevant authorities and pulls back what is likely to apply. Sub-second on a warm cache, with the source rail you can scroll line by line.
Validators check.
Every cite passes through multiple gates before the answer renders. Cases that fail verification are held back; the rest ship with the citation chip attached.
You ship the memo.
A clean seven-section memo with numbered citation chips, a scrollable source rail, and an audit-trail PDF. Ready for the matter file, ready for court.
Side by side
How Sanhita fits among the tools you already pay for.
Three honest comparisons. Numbers are sourced from each platform's public listings and reseller pages; figures shift quarterly and some incumbents only quote on enterprise contracts. Independent verification welcome.
Corpus depth, indexed records
Plenty of law to draw on
What the corpus actually covers
source mix
Price per seat
What you would pay this month
Monthly equivalents are the published annual rate divided by twelve. Most incumbents bill annually; only Indian Kanoon Pro, CaseMine, and the AI baselines (ChatGPT Plus, Claude Pro at the same tier) offer a real monthly option today.
One workspace vs. four tabs
Capability matrix
Sourced from each platform's public listings and recent industry reporting. Sanhita is still in closed beta; the rows where we show "partial" reflect the active gap-list (notably mature support / training and a public free tier). Independent verification welcome.
Impact
How the day actually shifts with Sanhita.
Three things start to change once Sanhita is in your stack: more time on the matter, less spent on tooling, and reach into the languages your clients actually speak. Numbers below reflect early pilot observations; we will republish as the cohort grows.
The market we serve. Concentrated, underserved, growing.
Four hours of judgment-hunting collapsed to under twelve minutes.
Replace four legacy subscriptions with one workspace.
Speak Hindi, Tamil, Bengali, Marathi. Sanhita listens.
Three matters before lunch, with citations clean enough to file.
Team
Two builders. One long road.
Sanhita is built and run by two technical founders who write every production line. No middle layer.
Ashutosh Singh
Platform, identity, India-region trust
Builds the systems that let every Sanhita user feel like the only one. Bank-grade identity. India-region data residency. Sub-second reliability across the surfaces that matter.
- Authentication and identity layer
- India-region cloud posture
- Trust, billing, observability
- Pilot onboarding for first advocates
Pranav Pandey
Knowledge, retrieval, drafting craft
Owns the depth and the discipline. Curates the corpus, builds the retrieval that makes ten minutes feel like four hours of research, and sets the quality bar every release must clear.
- Legal knowledge corpus and curation
- Retrieval and ranking
- Drafting library across 12 practice areas
- Evaluation against Indian benchmarks
Founding ML / RAG engineer. Founding product engineer.
Indian law data, real users from day one, equity meaningful enough to retire on if Sanhita works.
Get started
Two doors. Same workspace.
Have a code from your invite? Claim it. New here? Join the waitlist and we'll send you an access code once we've reviewed your application.