CiteFoundry Docs
Open dashboard

Quickstart

Sign up, create your first project, add a prompt, and read your first run — in about ten minutes.

This walkthrough takes you from a blank account to your first usable run. If you’d rather skim the model first, read Core concepts.

1. Create your account

Go to citefoundry.com and click Start free. You’ll sign in through WorkOS — Google, GitHub, or email all work. On first sign-in you’ll be dropped into onboarding to set up your organization (the billing-and-team boundary) and your first project (a brand to monitor).

2. Add a project

In the project picker on the top-left of the dashboard, pick + New project. Give it the brand’s display name and primary domain. The domain is how CiteFoundry attributes citations back to you across runs and how the brand-book is initially seeded.

You’ll land on the dashboard home. The empty state nudges you toward the next step: adding a prompt.

3. Add a prompt

A prompt is a question we’ll send to LLMs on a schedule. Go to Prompts → New prompt and write one. Pick something a real customer would type — not your brand name. Examples:

  • “What’s the best running shoe for marathon training?”
  • “Top project management tools for remote teams”
  • “How do I pick a CRM for a B2B startup?”

You can also bulk-import prompts later — paste a list, upload CSV, generate from a brief, or pull real queries from Google Search Console or Google Ads.

4. Attach the prompt to a monitor

Prompts on their own don’t run. A monitor binds a prompt to a model (ChatGPT, Claude, or Gemini) and a cadence (daily, weekly). Go to Monitors → New monitor, pick the prompt you just created, pick ChatGPT for the model, Daily for the cadence, and save.

The first run kicks off within a minute or two on Free; on paid plans it’s near-instant.

5. Read your first run

Once the run completes, open it from Runs. You’ll see:

  • The raw model answer.
  • Citations — every URL the model trusted as a source.
  • Brand mentions — who got named (yours and competitors’).
  • Surface metadata — model, run time, token usage.

If your brand wasn’t cited, the Content gaps view will already be collecting “prompts you’re losing” — that’s the lever you pull next.

Next steps