Curator Research
We build evidence-bound AI and knowledge infrastructure for Indian regulatory and document-heavy work: software, models, graphs, document intelligence and governed workflows designed around privacy, cost, evaluation and human authority.
01 / 03
Rules arrive as circulars, directions, Acts, judgments and filings — from many sources, in many formats, each changing on its own timetable. What matters is spread across documents that were never written to be read together.
02 / 03
We connect each source to what it amends, replaces, cites and depends on, with the date each link held. A question can then be answered from the structure — not from whichever page a search happens to find.
03 / 03
Our mark is a boundary closed on three sides and open on the fourth. Our systems show their sources and limits, state when evidence is insufficient, and keep consequential actions under human authority.
Connected tools for document work, knowledge, research, change and governed action.
For professionals
Patra — PDF tools, document management and knowledge-system integration, with a record of what was done to every file.
PDF tools built and in internal testing (31 operations, 25 Sep 2026) · document management — In development
For firms and regulated organisations
BanyanGraph, the connected law of India as a knowledge base; Curator, deep research with cited, date-correct answers; ChangeProof, a new rule traced through everything it touches.
Knowledge base in internal use (snapshot 27 Aug 2026) · public access, research and change tools — In development
For builders
Karta — a coding agent that works inside the limits a task allows, on a cost-conscious route, with a record of every step.
Internal evaluation on our own repositories — In development
We scope each engagement around the sources, permissions and decision to support. Access boundaries are tested before sensitive data is admitted. A company knowledge base service can connect approved organisational records into cited context.
Regulators are consolidating and rewriting large bodies of instructions. A policy that cites an older instrument can point somewhere that has moved.
An answer that cannot be traced to its source cannot be relied on in front of a regulator, an auditor or a court.
A useful answer names its source, respects effective dates and replacements, and says what it could not find.
Provenance, stated limits and human review are how the work is done — not a layer added afterwards.
We run the company on a governed fleet of AI agents. Every task has an owner, a limit, a reviewer and a record.
Every figure we publish carries its date and scope. Where a number is small, we say so, because the fraction is the honest part.
We model Indian law and regulation in its own terms: its instruments, its identifiers and the chain of authority that connects them.
Indian deep-tech for evidence-bound AI and knowledge systems.
Founded and directed by Bhavika Mukherjee and Aditi Chadha, Curator Research builds software, models, graphs and document intelligence for evidence-bound knowledge work, with privacy, cost, evaluation and human authority built into governed workflows. We plan to expand in phases, bringing in technical and functional specialists, collaborators and team members as products and engagements mature.
Tell us about the documents, the decision and the constraints. We start with scope and evidence, propose a solution, and then decide together what kind of engagement is useful.
