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

# Repository

> Your solution for finding contract data points from your existing agreements in bulk

Which contracts allow assignment without consent?  How many contracts include non-standard data use rights? What are the emails for security breach notices?

Find answers to questions like these across hundreds of contracts in minutes (not weeks).

<img src="https://mintcdn.com/dioptra-1e81a090/PyaFcE8u4unP5ldt/images/repository.png?fit=max&auto=format&n=PyaFcE8u4unP5ldt&q=85&s=0b3114bcad47a6e9b87f9e94b5950614" alt="Repository" width="1426" height="792" data-path="images/repository.png" />

### Creating an AI Column

You can ask for yes/no answers, actual extracted text, or summaries of issues.

<Tip>
  Here are some prompting instructions that you might find useful:

  * Extracting language:  "extract the complete language pertaining to....."
  * Returning yes/no responses: return the results as "Yes" or "No"
  * Summary: describe....
</Tip>

Here are some specific frequently used prompts:

* **Counterparty Identification**: The party to the agreement that is not "your own company name"
* **Document Type**: describe the type of document (e.g. email. master agreement, purchase order, amendment, etc)
* **Short Summary**: one sentence summary of the document
* **Parent Agreement Identification**: For documents that are purchase orders, orders, quotes, amendments, or other forms of documents that are related to (but are not) top-level agreements, identify the name and date of the controlling/top-level agreement (e.g. the master agreement, the software license agreement, etc).

## Where documents come from

Documents arrive in the repository two ways:

* **From a connected source** - imported automatically from a system such as your Icertis instance.
* **By manual upload.**

The settings panel lets you import, activate, and edit each source.

### Connecting an Icertis source

A connected Icertis source imports agreements automatically (on a schedule or on demand) and keeps them in sync.

<Steps>
  <Step title="Create a new source">
    Give the source a name (for example, `TFS Contracts`) and choose **Icertis** as the integration.
  </Step>

  <Step title="Scope the import with a filter">
    An import filter limits what comes across, so you are not syncing your entire instance. The filter runs against the Agreement entity. Examples:

    ```text theme={null}
    Name $like 'MSA*'
    Status $eq 'Executed'
    ModifiedDate $bt ['2026-01-01T00:00:00', '2026-06-14T00:00:00']
    ```

    Click **Preview** to see the live match count before anything imports (for example, *Filter matches 641 records*) along with the first few matching agreements and their entity types.
  </Step>

  <Step title="Set an import schedule (optional)">
    The schedule is a standard cron expression, in the order minute / hour / day / month / weekday. For example, `0 */6 * * *` runs every six hours.

    Leave the schedule empty for manual imports only.
  </Step>

  <Step title="Create the source">
    Click **Create Source**. The first import begins according to the schedule you set, or when you trigger it manually.
  </Step>
</Steps>

<Tip>
  Always run **Preview** before creating a source. A filter that is broader than you intended is much easier to catch at 641 records than after the import.
</Tip>

## Grouping agreements into families

Related agreements group into **families** so a master agreement and its amendments stay together instead of appearing as unrelated documents.

* Group by parent contract or by folder.
* For a connected Icertis source, the Icertis parent contract ID is used to build the family automatically.

<Info>
  Families matter for querying. A question asked across a repository of families returns the master and its amendments as one answer rather than several partial ones.
</Info>