intelligent.pageDocs

Extend intelligent.page

When to use directory settings, memory, connectors, skills, the Chrome companion, a second directory, projects and the API.

intelligent.page combines an organisation model with a task loop that reads, decides, builds and reconciles. Out of the box it reads your website and answers from what it learns. This page covers the layer you add on top: what you can teach it, connect to it, let it reach, and drive it from — and when each one is the right tool.

Note

For how the core loop works, see How intelligent.page works.

New here? Start by correcting the model under Memory and connecting one source of truth. Add the rest as specific triggers come up.

Overview

Each extension plugs into a different part of the loop:

  • Directory settings hold the standing facts every task starts from: the website, who you are answering for, the stage, the ontology
  • Memory is the model itself — editable, self-updating, and organised into sections — and the place a correction holds
  • Connectors let the work read the tools that hold the truth: repositories, calendars, analytics, trackers, chat, a marketplace account
  • Skills bring published thinking on the kind of problem you are working on into the run
  • The Chrome companion reaches what no server can — meetings, and pages behind your login — and can act in your browser on sites you allow
  • A second directory models a competitor, a client or an acquisition target with the same machinery, so comparisons are like for like
  • Projects give a measure you are trying to move a home, so tasks attach their milestones to it
  • Choice variables let one task hold alternatives open and return comparable variants
  • Agent personalities set the voice and stance each stage writes in, on the plans that offer them
  • Signals and subscriptions feed newsletters and other recurring material into the directory's inbox
  • Chat — Slack, Discord, Telegram, WhatsApp — lets you ask from where your team already talks
  • Search sources let the search bar query your own Slack, Drive, Jira, Notion and twenty more, live and as you
  • The API drives all of it from your own systems

Match features to your goal

Extensions range from standing context every task sees, to on-demand reach a task uses when it needs it, to places you talk to the directory from.

Build your setup over time

You do not need to configure everything up front. Each feature has a recognisable trigger, and most teams add them in roughly this order:

The same triggers tell you when to update what you already have. A repeated misunderstanding is a Memory edit, not a longer request. A stale website is an audit re-run, not a note in the composer.

Compare similar features

Some features can seem alike. Here is how to tell them apart.

Both hold standing context. Settings hold the few facts that frame everything; Memory holds everything else.

Change settings when the stage or the site changed. Edit Memory when a particular doc is wrong. Re-run the company audit when the site itself has moved on.

Combine features

Real setups combine them. A D2C brand might keep the stage current in settings, connect Shopify-side analytics and Seller Central, model two challengers in their own directories, run pricing tasks with three price points held open, and put trial-to-paid conversion in a project so every dossier's milestones land somewhere.

Understand the cost of each

Everything the work reads costs something, and so does everything it builds.

Learn more