Portal Mirror Construction

Overview

In order to construct portal mirrors - we need to retrieve all the data required from HubSpot.

We use two terms for this:

• Artifacts for large data exports, these encompass the commonly used Objects (Contacts, Companies, Deals etc.)
• Plugins for smaller exports, these are gathered quickly and are normally refreshed in full.

Artifacts and HubSpot Exports

Artifacts are large storage files of your HubSpot data - and they are split into two types:

1. The largest artifacts are created using the HubSpot export system, and are downloaded straight to the portal mirror.
2. Artifacts for smaller amounts of data (less common objects like products) are created using the HubSpot API on Daeda Servers and stored in Supabase - where they are then downloaded to the portal mirror.

This system allows us to build portal mirrors while minimising the amount of HubSpot data that passes through our servers. Currently Artifacts are rebuilt every 2 days to keep your portal data consistent.

Plugins

Plugins are smaller amounts of data that can be retrieved quickly and on demand - they include things like forms, workflows, lead scores etc. They are also split into two categories:

1. API Plugins - that use the HubSpot API.
2. Bridge Plugins - that require the Daeda AI Bridge to retrieve information.

Data retrieved by plugins is stored directly in the portal mirror, not in Supabase.

Refresh Mechanics

Keeping the data up to date is tricky - therefore we’ve exposed refresh tools to your AI system so it can trigger a data sync when it needs to.

It should do this automatically depending on the task at hand, but it might need to be reminded.

The following can be refreshed:

1. CRM objects since the last artifacts update.
2. API Plugins.
3. Bridge Plugins (if the user has access and has a chrome tab open to the portal).

The following cannot be refreshed:

1. CRM object artifacts (these currently refresh automatically every 48 hours)

If you are having issues with large amounts of CRM Objects, the issue is probably the artifact. Please contact support at support@daeda.tech

Fetch on demand

Artifacts deal with huge amounts of data, plugins deal with small amounts.

fetch_on_demand is the tool that handles large amounts of data that might only be useful in certain situations.

The current fetch_on_demand capabilities are:

1. List memberships.
2. Property Histories.

You or the AI can decide to fire off a fetch_on_demand to retrieve this information from HubSpot. Depending on the information it will take between a couple seconds to 2 minutes. You can ask the AI to “poll” and continuously check for when the information has been completed.