How VCinfo collects its data
VCinfo records venture investors and the companies they back from public sources, and keeps each source apart so a number can be traced to where it came from. This page says where the data comes from, how often it is read again and what it can miss.
Where the data comes from
- Investors' own websites. VCinfoBot reads each firm's public pages: portfolio, thesis, team and contact pages. A firm is published only when its own site shows that it invests, and its stated stages, sectors and regions come from what it writes about itself.
- Portfolio pages. The companies a firm lists are read from its portfolio page, including lists split over several pages. Each company is then checked at its own website: it is published only when that site answers and carries the company's name.
- Funding rounds. Rounds that name an investor come from Fundup AI. They confirm portfolio links, add companies a firm does not list, and show the stage at which it first invested. Round amounts in VCinfo Research posts come from Fundup AI as well.
- Public fund registers. Newly raised funds come from regulatory filings: SEC Form D in the US, private fund limited partnerships registered at UK Companies House, ESMA's EuVECA register, SEBI in India, MAS in Singapore, SPK in Turkey, CVM in Brazil and business.gov.au in Australia.
- News. Fund announcements from venture news feeds are published only once a second source confirms them. Headlines shown on profiles are matched to a firm or company with evidence and link to their outlet; many come through the GDELT Project.
- Public company registers. Facts on company pages, such as an incorporation date, come from sources like Companies House, SEC EDGAR and Wikidata, each shown with its source.
The crawler reads public pages only, follows robots.txt and does not try to get past bot protection. It does not collect personal email addresses or data about the people at firms. How VCinfoBot behaves.
How often it is updated
- Each investor's site is read again about every 30 days. Firms that are followed by users or are highly active are read about every 7 days.
- Funding rounds from Fundup AI, SEC Form D filings and Companies House registrations are checked throughout the day. The other fund registers are read about once a week.
- Fund news is read every few hours. Portfolio companies' own sites are checked again on a schedule, so a site that goes dark, moves or is parked is noted rather than left in place.
Every portfolio link carries the date it was last checked, on the site and in the API.
How changes are counted
A company counts as added when it appears on a portfolio page VCinfo already reads, and as removed only after two reads in a row miss it. A page that changes more than half of its list at once is held for a second read before anything is announced. A company leaving a portfolio page does not mean it closed or that the firm sold its stake.
What it can miss
- Not every fund or deal. VCinfo covers firms whose sites it can read and rounds that name an investor. Many firms list only part of their portfolio, some list none, and some rounds do not name their investors.
- Delay. A portfolio change is seen only when the firm updates its own page and VCinfo reads it again, so it can lag the deal by weeks or months.
- Round history. Funding rounds are complete from mid-2025. For a long-standing firm, round counts show recent activity, not its whole history.
- Amounts. VCinfo does not measure round sizes itself. Where a post gives an amount, it comes from Fundup AI and is only as complete as the round data.
- Registers. A fund missing from a register does not mean it was not raised: many funds are not required to file, or file somewhere VCinfo does not read.
How VCinfo Research uses this
Each post states its data source, the period it covers and how its numbers were computed, in a "Data & method" box at the top. Posts are drafted with AI assistance from VCinfo data and reviewed by the VCinfo team before they are published. If a number looks wrong, write to hello@vcinfo.io.
How each portfolio link is graded, from "declared" to "verified": Data sources and evidence grading.