Installing and Setting Up the Altify MCP Package

Install and set up of the Altify MCP package involves the following steps:

1. Install the package

2. Assign MCP permission sets to your admins and users

3. Configure your MCP Servers

4. Set daily token usage limits

Install the package

To install the Altify MCP package in your production or sandbox Salesforce org, do the following:

1. Click on the installation link provided by Altify and log in to the relevant org.

2. On the install page (shown below), select Install for Admins only and the check box to acknowledge that this is a Non-Salesforce application.

   

3. Click Install.

   Once the installation completes, you will see the confirmation screen showing 'Installation Complete!' with the package details.

Assign MCP permission sets to your admins and users

After installing the package, you need to assign the appropriate permission sets to your users. Two permission sets are included:

To assign the permission sets, do the following:

1. In Setup, find and select Permission Sets.

2. On the Permission Sets page, select one of the MCP permission sets listed above.

3. On the Permission Set page, click Manage Assignments - as shown in the example below.

   

4. On the Current Assignments page, click Add Assignment.

5. On the Select Users to Assign page, select the check boxes of the appropriate users and click Next.

   Note: A user who is both an administrator and an end user should be assigned both the Altify Remote MCP Admin and Altify Remote MCP User permission sets.

6. On the page Select an Expiration Option For Assigned Users, set a time limit to the permission set if required and click Assign.

7. On the Assignment Summary page, click Done.

8. Repeat steps 1 to 7 for each permission set.

Next, you need to configure the settings of the LLM that will be used by the MCP servers.

Configure your MCP Servers

The Altify MCP Configuration record stores the LLM provider settings used by the MCP servers. You need to create at least one configuration record.

1. In the Salesforce App Launcher, find and select Altify MCP Configurations.

2. On the Altify MCP Configurations page, click New.

3. On the New Altify MCP Configuration dialog, complete the fields as follows:

   • Configuration Name: Enter a name for this configuration (e.g. 'System')

   • LLM Type: The following LLM providers are supported. Enter the relevant value from the LLM Type column.

     LLM Type	Model Information	Service
     OpenAI	OpenAI Developers	OpenAI platform API
     Anthropic	Claude API Docs	Anthropic API
     Google	Gemini API	Google AI (Gemini) API
     Azure	Azure Foundry Models	Azure OpenAI Service
     Vertex	Overview of models on agent platform	Gemini Enterprise Agent Platform

   • Model: Enter the API model ID (e.g. 'gpt-5-mini'). Click the relevant link in the table above for further information.

   • User: Optionally assign this configuration to a specific user. If set, this configuration takes priority over the org-wide default for that user. Leave blank to create an org-wide configuration.

     This field is useful when different users need different LLM providers, models, or API keys.

   • API Key: Your API key for the chosen provider.

     Note: The API Key is stored securely within Salesforce. The MCP servers use these credentials to make LLM calls on behalf of users. Keep your API key confidential and rotate it periodically.

   • Customization: An optional JSON field for advanced settings such as temperature and token limits. See Advanced customization below.

4. Click Save.

Advanced customization

The Customization field accepts a JSON string for fine-tuning LLM behavior. This field is optional — sensible defaults are used when it is left blank.

The following options apply to all AI providers:

An example for most providers:

{"temperature": 0.1, "maxTokens": 2000}

An example for OpenAI GPT-5 series:

{"temperature": 1, "maxTokens": 2000}

Some providers require additional configuration fields in the Customization JSON:

An example for Azure:

{"azureInstanceName": "my-instance", "azureDeploymentName": "gpt-4o-mini", "azureApiVersion": "2024-02-01"}

Next, you need to confirm the MCP server usage limits for your users.

Set daily token usage limits

The Altify MCP package includes a custom setting called Altify MCP Usage Limits that controls token usage limits and logging behavior. You should review and update these settings as needed.

1. In Setup, go to Custom Settings.

2. Click Manage beside Altify MCP Usage Limits.

3. Click Edit for the 'Default' record to update the usage limit settings:

   • Enable Extended Logging: Standard logging always records which MCP servers and tools are invoked along with the associated token usage. If this setting is enabled, extended logging additionally captures the content of questions put to the MCP servers and their responses. This is for troubleshooting and audit purposes. Enable extended logging only when needed, as it increases storage usage.

   • Max Tokens Per Day: The maximum number of tokens that can be consumed per day across all MCP server calls (e.g. 1,000,000).

4. Click Save.

Altify MCP Server token usage

The following is how tokens are consumed by the servers:

• Token-tracked tools: Only tools that make LLM calls (such as TAS assessment analysis, vulnerability analysis, strategic recommendations, and other analytical tools) consume tokens and count towards the daily limit.

• Non-LLM tools: Tools that query or write Salesforce data directly (such as opportunity queries, data retrieval, and write operations) do not consume LLM tokens and are not subject to the daily token limit.

The Max Tokens Per Day limit is enforced on a per-day basis and resets daily. When the limit is reached, LLM-based tools will return an error until the next day. Non-LLM tools continue to work normally regardless of token usage.

Next, you need to create an ECA (External Client App) to securely integrate Altify's MCP servers with your AI assistant.