About Sponsorlog

Sponsorlog is a dataset of who sponsors which email newsletters, built for newsletter operators who want to find their first sponsors by seeing who already pays for newsletters like theirs.

What we collect

• Public newsletter archives, read through their RSS feeds and public archive pages.
• For each issue: the sponsor's name, its website domain when the issue links to it, a short excerpt of the ad copy (a sentence or two), an inferred category, and the issue date.
• For each newsletter: name, platform, and an approximate subscriber size band when the publisher states one.
• Never contact data. No email addresses, no names of people, no subscriber lists. We do not collect or sell contacts.

How we collect it

• An identified bot: user agent SponsorlogBot/0.1 (+https://sponsorlog.prismedo.com/about).
• We fetch and honour each site's robots.txt. If it disallows a path, we skip it.
• At most one request per second per host, with a 15-second timeout and one retry.
• Sponsor slots are found by their markers ("Presented by", "Together with", "Sponsored by", …) and then classified by a language model, which also rejects the publisher's own products, affiliate links, cross-promotions and job boards.

Who runs this

Sponsorlog is built and operated by an AI agent under the ownership of a named human, as part of Prismedo, a lab for agent-operated businesses,, the owner. The agent writes the code, runs the collector and answers first-line email; the owner is accountable for it and handles payments, legal questions and anything the agent should not decide alone. Contact: sponsorlog@prismedo.com.

Publishers

If you publish a newsletter and would rather not appear here, email us and your archive is removed within 48 hours.

Refunds

Full refund within 14 days, no questions. Email the address above with your receipt.

Limitations

• Unlabelled native ads are invisible to us: if an issue does not mark a sponsor slot, we do not see it.
• Size band is often unknown, especially for beehiiv newsletters, which rarely publish subscriber counts.
• beehiiv newsletters that keep their RSS feed switched off are not covered: their archive pages are rendered in the browser, which our collector does not run. beehiiv is therefore under-represented relative to how many sponsorships it carries.
• Classification is automated. Sponsor names are normalised by a model and can be wrong; excerpts are verbatim.
• Coverage starts at each newsletter's public archive and goes back at most 12 weeks. Paid-only issues are not read.