This page describes how to upload your mainframe application files to
Mainframe Assessment Tool and create an assessment.

You can perform an asset discovery on your own, or request
assistance from a Google Customer Engineer.

## Before you begin

To prepare for your assessment, make sure that you have completed the following
prerequisites:

1. [Configure your Google Cloud project for Mainframe Assessment Tool](https://docs.cloud.google.com/mainframe-assessment-tool/docs/configure-your-project.md.txt).
2. [Set up and access Mainframe Assessment Tool](https://docs.cloud.google.com/mainframe-assessment-tool/docs/create-instance.md.txt).
3. Create a zip file containing all the applications that you want to migrate:
   - Mainframe Assessment Tool supports only zip files with a maximum size of 400 MB. If your application zip file is larger than 400 MB, consider splitting the file into smaller zip files based on business domains and uploading a zip file for each domain.
   - If the applications contain references to external artifacts such as programs or tables, we recommend that you include these dependencies as well. For the list of supported languages, databases, transaction monitors, and scheduler configurations, see [Supported languages](https://docs.cloud.google.com/mainframe-assessment-tool/docs/overview.md.txt#supported-languages).
   - Don't include binaries and other non-source files in the zip file.

## Create an assessment

To upload the zip file and create an assessment in Mainframe Assessment Tool, follow
these steps:

**Assessment details**

1. To open Mainframe Assessment Tool on your local machine, open a web browser window
   and go to `http://localhost:PORT`.

   Replace `PORT` with the port number you used on your
   local machine for the tunneled connection to Mainframe Assessment Tool.
2. On the **Home** page, click **Create assessment**.

3. For **Name**, enter a name for the assessment.

4. Optional: For **Description**, enter a description for the assessment.

5. To upload a zip file, click **Select file**. The file browser opens. Select
   a zip file that you want to upload. You can upload only one file at a time.

6. Click **Next**.

**AI insights**

1. Optional: To enable AI insights for the assessment, select **Enable AI
   insights**.
2. To generate paragraph-level summaries for COBOL code, select **Detailed
   summaries**. By default, this option is selected.
3. To generate test cases for the assessment specifications generated by Gemini, select **Test cases**. By default, this option isn't selected.
4. To extract information about databases and files accessed by the
   application, select **Data sources**.

   > [!NOTE]
   > **Note:** Enabling **Detailed summaries** , **Test cases** , or **Data sources** for zip files larger than 5 MB is discouraged because it can increase processing time and cost.

5. Optional: For **Additional generation guidance**, enter free-text guidance
   to add to the Gemini prompts during AI generation.

6. Click **Next**.

**Model \& quality**

Configure the Gemini models for the assessment. These settings
apply only to this assessment.

1. Select how Mainframe Assessment Tool routes assessment steps to Gemini
   models:

   - **Fast**: returns the quickest results by using lighter models.
   - **Balanced** (recommended): balances speed and depth for most assessments by routing complex reasoning steps to heavier models and simpler steps to lighter models. This is the default option.
   - **Extended**: maximizes analysis depth by using heavier models.

     > [!NOTE]
     > **Note:** The **Extended** level uses heavier models and can significantly increase costs, especially for large assessments.

   - **Use a specific model instead** : runs all assessment steps on a single
     model rather than optimizing by level. Select a model from the list.
     Each model is labeled either [**General
     Availability**](https://docs.cloud.google.com/products#product-launch-stages)
     or [**Preview**](https://docs.cloud.google.com/products#product-launch-stages).

   Each quality level lists the Gemini models used for each
   step. Mainframe Assessment Tool supports all zones within the available
   [Agent Platform
   regions](https://docs.cloud.google.com/vertex-ai/generative-ai/docs/learn/locations#available-regions).
2. Optional: To restrict model calls to the assessment region and its
   associated multi-region, enable **Keep data in my region**.

   By default, Mainframe Assessment Tool routes model calls to the optimal location for
   the selected models, including the assessment region, a multi-region, or the
   global Agent Platform endpoint. When this
   setting is enabled, Mainframe Assessment Tool performs the following actions:
   - Disables quality levels that require a Gemini model that isn't available in the assessment region. These levels are labeled **Not available in this region**.
   - Hides models from the specific-model list that aren't available in the assessment region.

   If no supported model is available in the assessment region, the assessment
   can't run with **Keep data in my region** enabled. To proceed with the
   assessment, disable **Keep data in my region** to allow model calls outside
   the region.

   If your Mainframe Assessment Tool instance doesn't have an external IP address and
   you enable **Keep data in my region** , models served only on a multi-region
   endpoint might be unavailable. This happens because
   [Private Google Access](https://docs.cloud.google.com/vpc/docs/private-google-access.md.txt) can't reach
   multi-region endpoints. To make these models available while adhering to
   data residency requirements, set up private connectivity to the multi-region
   endpoint. For instructions, see [Configure private connectivity for data
   residency](https://docs.cloud.google.com/mainframe-assessment-tool/docs/configure-your-project#private-connectivity).
3. Click **Next**.

**Optional: Domains**

1. Optional: To add one or more domains that represent a distinct area of
   functionality, click **Add domain**.

   Domains let you filter and review results based on specific contexts after
   the assessment is complete.
2. Click **Next**.

**Create assessment**

1. Click **Create assessment**.

After you upload the zip file, Mainframe Assessment Tool unpacks and processes
the source code. The **Assessments** page displays a new entry in the table
for each discovered asset, showing its assessment status:

- **Pending:** assessment is waiting its turn.
- **In Progress:** analysis is in progress.
- **Analyzed:** assessment has been completed.
- **Failed:** the assessment didn't complete.

If you see errors after importing your file, you can review the error message
directly in Mainframe Assessment Tool.

To view more details about the assessment progress for an item, click the
reference ID next to it in the table. This opens a page that shows all the
steps together with the status.

## Add files to an existing assessment

After you create an assessment, you can add additional zip files to it.
The files from the new zip file are merged with the existing files in the assessment.
If you want to **replace** existing files, the new zip file must include the
files that you're replacing with the exact same path structure.

To add files to an existing assessment, follow these steps:

1. Open Mainframe Assessment Tool in a web browser.
2. In the navigation menu, click **Assessments**.
3. On the **Assessments** page, find the assessment that you want to modify, and click **Open assessment**.
4. Click **Add a file**.
5. Click **Select file**, and select a zip file that you want to upload. You can upload only one file at a time.
6. Click **Add**.

After you upload the files, Mainframe Assessment Tool starts the analysis again.
If you replace files, they are fully reassessed, and any existing analysis
for these files might change. During reprocessing, a **Stale** indicator
appears next to the files that are being reprocessed.

## Update assessment details and settings

After you create an assessment, you can update its name and description,
and modify AI-generated insights settings.

To update assessment details or settings, follow these steps:

1. Open Mainframe Assessment Tool in a web browser.
2. In the navigation menu, click **Assessments**.
3. On the **Assessments** page, find the assessment that you want to modify, and click **Open assessment**.
4. Click settings **Edit assessment settings**.
5. Optional: On the **Assessment details** step, edit the name and description and then click **Next**.
6. Optional: On the **AI insights** step, modify AI settings as needed, and then click **Next**.
7. Optional: On the **Model \& quality** step, configure the models or quality level, and then click **Next**.
8. On the **Domains (optional)** step, click **Next**. You can't edit domains after you create an assessment.
9. Click **Update assessment**.

After you update the settings, Mainframe Assessment Tool starts the analysis again.

## What's next

- Learn how to [Review and understand assessment results](https://docs.cloud.google.com/mainframe-assessment-tool/docs/review-results.md.txt).