Projects
Settings
Project identity, included crawl hosts and paths, autonomy rules, business context, report cadence, and archiving.
Project Settings defines who the site is for, what the engine may touch, and how much it may do on its own. Choose Save settings after editing; the Autonomy section has its own Save autonomy button.
Project
Name, Domain, and Canonical site URL identify the site. Turn on Scheduled authority evidence to allow weekly paid backlink and referring-domain refreshes when the autonomy policy permits them and credits are available. New plans do not have a separate project dollar-budget field; manage the account's credit balance and recharge policy in Settings → Plan & usage. Existing Core subscriptions still show their legacy monthly evidence budget.
Autonomy and safety
Choose how much the engine owns:
- Audit only / MCP — observe and file findings only. You or an external agent handle every change.
- Copilot — execute low-risk work; ask before risky, external, or paid actions. This is the default for new projects.
- Full — execute policy-allowed work autonomously, including paid evidence while credits are available.
- Custom — set a default rule, then override individual departments or action classes.
Every rule has the same controls: Create tasks, Prepare plans, Execute, Maximum risk, Spend limit (USD) per prepared action, and Require approval. Custom lets you give each department a narrower or broader boundary and pin rules to concrete action classes such as Site crawl, Metadata update, Directory submission, Authority refresh, Task comment, and Task status update. Project scope and pre-execution checks still apply in every mode. See Autonomy for how approval and risk work.
Business context
Business description, Audience, Conversion goals, Brand constraints, and Competitor domains tell the engine what matters. It uses them to prioritize work, name topic clusters, judge competitors in AI answers, and constrain any content it drafts.
Scope
Included hosts takes one host per line, such as www.example.com or app.example.com. Only the project domain and its subdomains can be crawled. Search Console and sitemaps may reveal other hosts, but they are not included automatically. Included paths and Excluded paths take one prefix per line, for example /blog or /checkout. Blank included paths mean every path on the included hosts; exclusions always win. Markets and Locales describe where the site competes.
Reports
Turn on Scheduled reports, choose a Weekly or Monthly cadence, and set the Timezone the periods should follow. Email and channel delivery are not currently active. See Reports.
Goals
The project goal is edited from Active goals beside Tasks, not from Settings. Changing the goal reprioritizes findings and future tasks; it never changes autonomy, integrations, or spend limits.
Integrations
Integrations are connected once at the workspace level. When a Search Console property, GA4 property, or Cloudflare zone matches the project domain it is attached automatically; see Integrations.
Archiving
Archive project stops all engine work and hides the project from lists. Stored evidence is kept. Archiving does not change billing; the paid site position stays available for another project.
Setting changes affect future planning and actions. A plan that is already frozen keeps its recorded policy and target until it is rejected or prepared again.
