Skip to main content
This page takes you from signing up to your first results: an organization for your team, a finished setup, an agent readiness score for your website, and the first runs of your prompts and evaluations. Plan for setup to run in the background while you do other things.
Agentrail needs a browser window at least 1024 pixels wide.

​
Sign up and set up

1

Create your account

Go to app.agentrail.co and select Create an account. On Get started with Agentrail, select Continue to sign up and finish on the secure sign-up page. Agentrail opens when your account is ready.
2

Create your organization

On Create your organization, enter an Organization name and your product’s public Website, such as example.com. Select Create organization. You become its Admin, and setup opens.Joining a teammate’s organization instead? Open the link in your invitation email. See Accept an invitation.
3

Start setup

Under Research your product, select Start setup. Setup reads your website, fills in the Brand Vault and proposes competitors. It runs in the background, so you can leave the page; progress is kept.
4

Review what research found

Check the Brand Vault and Competitors tabs and fix anything wrong or missing. Drafting uses both exactly as you leave them. Then select Continue to drafting.
5

Review setup and finish

Setup drafts topics, preference prompts and experience evaluations. Check the Topics, Preference prompts and Experience evaluations tabs. Turn on the Needs attention filter to find anything missing a topic or criteria. Then select Finish setup.Finishing starts running your prompts and evaluations in the background, and Agentrail opens.
Only an Admin moves setup forward. If setup stops with Research could not finish. or Drafting could not finish., select Try again. The full walkthrough is in Setup.

​
Run your first scan

Agent readiness checks whether AI agents can get into, read and use your public website, documentation and API. Any member can start a scan.
1

Open Agent readiness

In the sidebar, select Agent readiness.
2

Scan

Select Scan now. The scan appears under Active scan.
3

Read the score

When the scan finishes, Latest score shows a score out of 100 and a grade. Open the scan to see every check, what it found and how to fix it.
After the first scan, scans start on their own at the interval set in Settings on the Agent readiness page.
To scan your docs or app as well as your website, an Admin adds them under Documentation and application URLs in Settings. Saving with readiness on starts a scan.
See Agent readiness to understand the score and the checks.

​
See your first runs

1

Watch the runs

In the sidebar, select Executions. The runs started by finishing setup appear as soon as they start, with their status. Open one to follow its log.
2

Read the results

As runs finish and are judged, Agent Preference shows which products coding agents discover, recommend and choose for your prompts. Agent Experience shows how far agents get when they build with your product, and what stops them.
3

Run a prompt yourself

An Admin can start more runs at any time: select Run on a prompt or an evaluation. It starts one run for each of its agents and models.

​
Next steps

Invite your team

Add people as Admins or Members once setup is finished.

Core concepts

How organizations, workspaces and the parts of Agentrail fit together.

Agent Preference

Read the rankings, head-to-heads and reasons agents give.

Agent Experience

See how far each task got and where agents failed.