Skip to main content
A project is a single audit context: one brand, in one market, with the set of URLs you want measured together. It’s defined by a project name, a primary domain, and the language and country you’re auditing for. Those settings shape how personas and prompts are generated. All URLs in a project share the same personas, prompts, snapshot schedule, and visibility metrics, so everything inside a project is measured against the same audience and the same competitive landscape.

What a project contains

Each project has its own:
  • URLs: the pages you want audited
  • Personas: who your pages are written for, extracted from your content
  • Prompts: audit prompts generated from your personas and page content
  • Snapshots: weekly re-runs of the full audit pipeline
  • Recommendations: prioritized content and technical fixes per URL and prompt

When to create a new project

You can run as many projects as you have slots for, and each one runs independently: its own URLs, personas, prompts, and snapshot schedule. Create a separate project whenever the audit context genuinely differs:
  • A different brand or domain: each brand gets its own personas, competitors, and metrics.
  • A different market: the same brand audited in another country or language. Personas and prompts are generated per language and country, so a German audit and an English audit belong in separate projects.
  • A distinct product line or site section that serves a different audience and would otherwise dilute the personas.
Keep a single project when the URLs share the same audience and market. Splitting them up fragments the topic clusters and cross-URL analysis that let your pages reinforce each other.
Rule of thumb: one brand, one market, one project. Reach for a new project when the audience or the language/country changes, not just because you have more URLs.

Team members

A project can be shared with unlimited team members. Invite teammates by email from the project’s settings, and they get access to that specific project. Each account can send a limited number of invitations per day. Members hold one of three roles:
  • Owner: full control of the project, including inviting and removing members and archiving it. Every project keeps at least one owner.
  • Editor: can run audits, manage pages, review personas and prompts, and work with the results.
  • Viewer: read-only access to the audit and its results.
Seats are scoped to a single project. Inviting someone to one project doesn’t give them access to your other projects.

Slots

Projects live in slots. Each slot is a licensed capacity unit tied to your plan. You can have slots without projects attached, ready to assign to a project. A slot stays with its project until you archive that project, as described below. Every account includes a free demo slot, which is a scoped trial version of the full product. Each paid slot you add also unlocks an additional demo slot, so agencies and teams can explore more brands alongside their active client work.

Archiving a project

Archiving removes a project from your active workspace while preserving its slot.
  • If nothing has been analysed yet: archiving is immediate. A project that has not crawled a page or collected a single answer has nothing to keep, so it is removed and its slot becomes available right away. This is the path for a project you set up with the wrong domain.
  • Once analysis has run: the project is restorable for 30 days after archiving, and everything in it is kept for that whole period. Its slot stays with it during that time.
  • Past the 30 days: the project and its analysis data are permanently deleted, and the slot becomes available again.
Restoring an archived project also returns its slot to use straight away, so if you archived something by mistake, or you need the slot back before the 30 days are up, restore it rather than waiting.
The slot staying with the project is deliberate. Archiving is not a way to reset a slot and start over on a fresh brand: analysing a project costs real work on our side, so a slot you have already used comes back on the same schedule as the data it produced, not sooner.