ContinuumBot
Who we are, what we read and why, what we keep, how we behave, and exactly how to stop us.
Who we are
ContinuumBot is the crawler of Continuum Alternatives, private capital data with the source behind every number. Every request it makes carries this user agent:
ContinuumBot/1.0 (+https://continuumalternatives.com/bot)Every request also carries a From: header with the contact address below, so the person responsible for the crawler can be reached from any server log. The identity is never disguised — not to pass a robots file, not to pass anything.
What we read and why
We read publications made for the professional or public record: public registers and regulators, organisations’ own websites, filings, conference programmes and the press. From them we record facts about organisations, funds, deals and the professional roles people hold.
We never:
- read anything behind a login or a paywall;
- pass a CAPTCHA, a bot challenge or an IP block, by any means;
- accept terms of use by a click, or create an account in order to read;
- read LinkedIn, or take anything about a person from a professional network beyond a link the firm itself publishes;
- copy the databases of data sellers.
What we keep
We keep facts, each with the source and the place it came from. A page is kept only as evidence for the facts drawn from it; it is never shown or served to anyone. Where a site reserves its content from mining or AI use by a machine-readable signal, we keep no copy at all and extract facts by code only. We do not republish others’ text or images.
About a person we hold only the professional record: their name, the roles they hold or held with their dates, the professional headline as the organisation publishes it, the country of the office a role is based in, the identifiers a register assigns to their public professional record, what the firm itself publishes about them on its own site, and the shareholdings or control a public register publishes for them, as a seat. Never identity numbers, home addresses, dates of birth or personal contacts.
How we behave
- One connection per host at a time.
- At least 5 seconds between requests to a website, or the robots file’s
Crawl-delayif that is longer. Retry-Afteris obeyed; on 429 and 5xx responses we back off exponentially, and repeated errors pause the host for 24 hours.- Conditional requests (
If-None-Match,If-Modified-Since), so an unchanged page costs nothing. - Bulk work at the host’s night; sitemaps and feeds before walking links.
- Nothing is clicked except to dismiss or accept a banner that asks only about cookies. A banner or page that accepts terms of use by a click ends the request.
How to stop us
One way to stop us outright, two ways to limit us; a request to us works in every case.
1. Name us in your robots file
A rule naming ContinuumBot stops us on every channel. Add these two lines to /robots.txt:
User-agent: ContinuumBot
Disallow: /Plainly: a rule for all crawlers alone (User-agent: *) does not stop us reading organisation facts, nor reading an API under its terms or a file published under an open licence (there the rule is recorded, not applied). On any other page it limits us to facts only — no copy of the page kept, no personal data — at half pace. To stop us entirely, name ContinuumBot.
2. Use a machine-readable AI or mining signal
We read and honour these, and any other machine-readable opt-out from text and data mining:
Content-Usage— the AIPREF header or robots ruleContent-Signal— the robots file linetdm-reservation— TDMRep: the header, page metadata or /.well-known/tdmrep.jsonai.txt— the file at the site rootnoai, noimageai— page metadataX-Robots-Tag: noai— the response headerLicense (RSL)— the robots file line, where it prohibits AI input or all use
Their effect: facts only, no copy of the page kept, and no language model reads the text — extraction by deterministic code only. The URL stays marked for every later visit.
3. Ask us
Write to bot@continuumalternatives.com. A request to stop reading a site, or to remove content we hold from it, pauses the source at once; a person then decides and answers you. For a request about a person — access, correction, erasure or objection — write to the same address; a request form will follow.
Our outgoing addresses
Our published IP ranges, with matching reverse DNS, will be listed here once the crawler has its fixed outgoing address.