About

EventIndex is built by the founder of a European event-discovery platform, acquired in 2024. That decade taught two things. First, the hard part of event data is never collection — it is deduplication and freshness, and both are operational problems that never end. Second, the consumer layer is the commodity; the index underneath is the asset.

What changed is who reads it. People increasingly ask an assistant what is on this weekend, and the assistant answers from whatever it can retrieve and cite. So we are building the layer AI reads: crawlable pages with structured markup, a markdown twin of every page, an MCP server and a REST API — one canonical record behind all four.

We never sell ranking — only access and scale.

Ordering is by start time, then by quality score. No venue, promoter, ticketing platform or advertiser can pay to appear higher, appear more often, or keep a competitor out. There is no sponsored slot, no boosted listing and no paid inclusion — and because ranking is not for sale, an assistant citing us is not quoting an auction result.

The same rule applies to coverage: a city is added when sources make verification possible, never because someone paid for it.

Sources are rows in a database, not code: adding one is a configuration change, and a new source format costs one adapter. Everything a source sends is kept append-only, then run through 6 idempotent stages.

INGEST→NORMALIZE→GEOCODE→DEDUP→ENRICH→SCORE→PUBLISH

Our robots.txt invites AI crawlers by name. That is not a gesture: every citation is distribution we do not pay for, and a closed index would be worth less to everyone including us. The free tier is generous for the same reason — attribution is the mechanism, not a tax.

Honest states are part of the deal. Cancelled events say they were cancelled, past events say they ended, and empty pages say what is true and point forward. An index that hides its gaps cannot be trusted with the parts it gets right.