> ## Documentation Index
> Fetch the complete documentation index at: https://docs.rippit.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How data integrations work

> How Rippit brings your conversation data in, what it stores, and how to disconnect.

A data integration is a read-only connection to a system that holds your customer conversations. Rippit pulls those conversations in on a schedule and stores them as tables you can query, build reports from, and point Agent Apps at.

Data integrations and [MCP servers](/agentic-chat/overview) do different jobs. Integrations bring your conversation data into Rippit in bulk. MCP servers unlock the ability to read, write, and update data in other systems on an ad hoc basis.

This page covers what is true of every data integration. Each integration also has its own page for the details specific to it.

## Why Rippit ingests rather than queries live

Analysis needs the whole corpus at once. Questions like "what changed in billing complaints this quarter" or "which issues have no help article behind them" read across thousands of conversations, and no helpdesk API is built to answer that in a single request.

So Rippit copies the conversations in, keeps them current, and does the analysis against its own store. The practical consequences:

* **Results are fast**, because nothing round-trips to your helpdesk at query time.
* **Your helpdesk is not under analysis load.** Rippit's API usage is a steady background sync, not a spike whenever someone asks a question.
* **Data is a few hours behind, not live.** See the individual integration page for its sync behavior.

## What Rippit stores

Full conversation records: the message history, the participants, and the fields attached to each conversation, including custom fields.

**There is no storage limit.** Rippit does not cap how many conversations you can bring in, and adding more history does not cost you anything in storage.

The first sync backfills up to 90 days of history. After that, new and updated conversations continue to sync in.

## Access model

Every data integration is **read-only**. Rippit authorizes through OAuth, so no password or API key is shared, and nothing Rippit does writes back to your source system.

Rippit can read only what the account that authorized the connection can read. Use a dedicated service account where the source system supports one: you control what it can see, and the connection does not break when someone leaves.

For the full security model, see [Data integration security](/security/data-integrations).

## Supported integrations

Rippit currently ingests from Intercom and Zendesk. Each has its own page covering authorization, the API endpoints it calls, and its sync behavior.

## Disconnecting

**Disconnect** on an integration's page stops future syncing. Revoking Rippit's authorization in the source system has the same effect, since the token stops working immediately. Neither deletes conversations Rippit has already imported.

**Delete account**, on the Billing page, ends the account at the end of the current billing period or immediately. It is a credential and access teardown: all users are deactivated, sessions and API credentials are revoked, integration authorizations are marked dormant, and stored third-party credentials are deleted.

It deletes no ingested data. Conversations, message content, the tables built from them, and their dashboards and column configuration are all retained.