india-v2-bot
An automated client that requests published documents from Indian financial market publishers. If it has reached your site, this page is what it is.
What it sends
india-v2-bot/1.0 (+https://raketaai.github.io/india-v2-bot/)
That is the identity it sends everywhere but at two publishers, and everywhere else it never
asserts a Referer header.
The two exceptions are the National Stock Exchange of India and BSE. Both refuse
any client that does not present as a web browser: measured on 23 September 2026, BSE refused this
name on every request and NSE left every connection unanswered. To those two alone the program sends
the headers of a Chrome browser, with the Referer of the exchange's own site, and to
BSE also the Origin of BSE's own site. Nothing else is added. It asks them
only for filings the Listing Regulations make public, it honours their robots.txt —
including a rule that names india-v2-bot, though it does not send that name there — and
every such request is recorded against this exception. If either exchange would rather admit the
program under its own name, write to the address below and it will switch.
What it is
It collects published documents — issuers' financial results, annual reports and regulatory disclosures, credit rating agencies' rating rationales, and the lists of filings and ratings that index them — from exchanges, regulators, credit rating agencies and issuers in India. They feed a private research platform on the issuers of the rupee bond market. Nothing collected is redistributed, resold, licensed or served to anyone.
It is automated, and its code is written by AI coding agents (Anthropic's Claude), under the operator's direction and review. That is disclosed here rather than compressed into a token, and it is not a claim to act for Anthropic: this program is not one of their crawlers and does not send their crawler names.
Who operates it
Andrey Fomin, an individual, on his own account. He acts for no company,
client or other principal, and no third party is a party to this. The domain raketa.ai
and the GitHub account RaketaAi are his own. Requests leave from two hosts he
operates: one in Mumbai, India — currently 65.20.72.152 — and one in Germany —
currently 178.104.188.64. The second asks again only for what the first could not
obtain.
How it behaves
- It honours your
robots.txt, including where doing so costs it documents it wants. It follows RFC 9309: a rule namingindia-v2-botis obeyed wherever it is written, and where no rule names it, the rules for every robot apply. If yourrobots.txtis missing or refused, nothing is disallowed; if it does not answer, the program asks for nothing on your site until it does. - It asks for specific things: an issuer's list of filings or ratings, found through the site's own search, and the documents that list names. To learn how a site lists them it may fetch the site's front page and scripts once. It does not follow links to see what a site holds.
- It is slow on purpose: at most one request every two seconds to any host.
- It never creates an account, registers, accepts terms, asserts a membership, or pays for anything — in anyone's name, including the operator's. A source behind a login or a membership is recorded as not asked, and left so.
- It does not disguise itself, except as stated above. Anywhere else, if you refuse this name, it is refused: the refusal is written down and the source is done without, rather than worked around.
How to make it stop
Refuse it in your robots.txt and it will stop. It reads the file
and obeys it:
User-agent: india-v2-bot Disallow: /
Or say so directly — write to contact@raketa.ai, or open an issue here. Name the host and, if you have it, a log line. It will be stopped, and no reason is needed.
This page's own history
This page is served from a public git repository, so what it said on any given day — and what this program therefore claimed to be at the time it made a request — is readable at github.com/RaketaAi/india-v2-bot rather than being whatever it says today.