The reader that opens your website when you ask us to.

What it is

The program Envoyra uses to read a public website when a person asks us to look at it. Someone types in their own address, or one we are tracking for them, and the reader opens the pages a visitor could open.

What it fetches

A handful of public pages from the site the person named, to learn what the business sells so we can suggest the questions to ask AI assistants. It also opens pages that an AI assistant cited in an answer we recorded, to report what kind of page it is. It reads the page as a visitor would. It does not sign in, fill in a form, buy anything or click through a checkout.

What it does not do

It is not used to train AI models. It does not copy a site to republish it. It does not run all day: it reads when a person asks, a few pages at a time.

How it identifies itself

Every request carries this user agent. We never present as an ordinary web browser or under any other name:
EnvoyraReader/1.0 (+https://www.envoyra.com/envoyra-reader)

It follows robots.txt

Before reading a site it checks that site’s robots.txt and stops if the file turns it away. To allow it, add:
User-agent: EnvoyraReader
Allow: /

If a firewall is blocking it

Robots.txt is not the only gate. Cloudflare, Akamai, Fastly, AWS WAF and similar services can turn automated readers away before they see robots.txt. In your bot or firewall settings, allow requests whose user agent contains EnvoyraReader. Envoyra reads from shared cloud addresses, so there is no fixed list of addresses to allow; the user agent is the identifier.

Why you might want to allow it

A site that turns our reader away often turns away other automated readers too, and AI assistants are among them. If you are reading this because Envoyra reported that your website could not be read, it is worth checking what else your settings block.

Contact

Questions, or a request to stop reading a site: hello@envoyra.com.