Start
How Webtick works
A website growth operator: it observes, ranks opportunities by evidence, prepares changes, and leaves merge and publish to a person.
Webtick runs one loop over your website: observe, understand, discover, design, execute, measure and learn. It works through the tools you already use and never asks you to migrate.
| Step | What Webtick does | Where you see it |
|---|---|---|
| Observe | Audits the site; syncs Search Console, GA4 and other sources | Intelligence, audits |
| Understand | Keeps a confirmed business profile and competitor set | Project → Profile |
| Discover | Turns evidence into ranked opportunities | Opportunities |
| Design | Turns an opportunity into a scoped change proposal | Proposals |
| Execute | Opens a GitHub pull request or saves a CMS draft | Proposals, Activity |
| Measure | Tracks keywords, AI answers and page metrics over time | Tracking, Intelligence |
| Learn | Keeps decisions and outcomes on the record | Activity, decisions |
What Webtick never does
- Never merges a pull request. A pull request is the only route for code.
- Never publishes CMS content. It saves drafts, with your CMS's own version history.
- Never gives a credential to an AI model. A model receives text and returns text; the worker does every write.
- Never changes your hosting, DNS or infrastructure, and is never your CMS, site builder or host.
- Never invents claims: content uses only statements you confirmed in the business profile.
The parts of the system
| Part | Role |
|---|---|
| Web app (use.webtick.app) | The product UI |
| API (use.webtick.app/api/v1) | Every product action; OpenAPI at /api/v1/openapi.json |
| Worker | Crawls, audits, syncs, research, AI drafting and execution, as background jobs |
| Model gateway | The only path to a language model: text in, text out, with a budget per call |
| Database | PostgreSQL with row-level security per workspace |
Next: Workspaces and projects