EphemeralBot

Recognising its requests

Every request carries this user agent:

Mozilla/5.0 (compatible; EphemeralBot/1.0; +https://ephemer.al/bot)

It also carries an Ephemeral-Request-Id header, unique to the request. Quote it in a report and we can find the run that sent it.

Requests come from this address: 35.72.133.50. The current list is also at /bot/ips.json. These addresses send nothing but EphemeralBot requests.

How it behaves

• Only runs whose owner turned web access on can use it, and an owner can limit a run to a list of domains.
• Each request is checked before it is sent. Requests that look like attacks, spam or someone smuggling data out are refused.
• All customers together send a site at most 60 requests a minute, and each run at most 50 in total.
• It never reaches private or internal addresses, and follows at most five redirects, checking each one.

robots.txt

EphemeralBot follows the rules a robots.txt file addresses to it by name. It does not follow rules written for every crawler (User-agent: *), because each request is made for a task someone gave it, and those rules often block APIs that people rely on. To keep it out of all or part of your site:

User-agent: EphemeralBot
Disallow: /

Changes take effect within an hour.

Reporting abuse

Report it below with the Ephemeral-Request-Id of a request, or its time, URL and source address. We can identify the customer behind every request, and we stop accounts that use EphemeralBot for attacks, spam or scraping personal data.

Prefer email? Show our abuse address

Building with Ephemeral? Web access in the docs explains how to turn it on for a run.