Skip to main content

Braze MCP Server

Braze Model Context Protocol (MCP) server is a secure connection that lets AI tools like Claude and Cursor access non-PII Braze data to answer questions, analyze trends, and provide insights. It is primarily read-focused and also supports specific write actions such as media library asset creation when configured. For more general information, see Braze MCP server.

Additional resources

Quick start

Step 1: Create an API key

In Braze, go to Settings > APIs and Identifiers > API Keys and create a new key with some or all the following permissions.

Category Endpoint Required Permission
Campaigns /campaigns/data_series campaigns.data_series
Campaigns /campaigns/details campaigns.details
Campaigns /campaigns/list campaigns.list
Campaigns /sends/data_series sends.data_series
Canvas /canvas/data_series canvas.data_series
Canvas /canvas/data_summary canvas.data_summary
Canvas /canvas/details canvas.details
Canvas /canvas/list canvas.list
Catalogs /catalogs catalogs.get
Catalogs /catalogs/{catalog_name}/items catalogs.get_items
Catalogs /catalogs/{catalog_name}/items/{item_id} catalogs.get_item
Cloud Data Ingestion /cdi/integrations cdi.integration_list
Cloud Data Ingestion /cdi/integrations/{integration_id}/job_sync_status cdi.integration_job_status
Content Blocks /content_blocks/list content_blocks.list
Content Blocks /content_blocks/info content_blocks.info
Content Blocks (write) /content_blocks/create content_blocks.create
Content Blocks (write) /content_blocks/update content_blocks.update
Custom Attributes /custom_attributes custom_attributes.get
Events /events/list events.list
Events /events/data_series events.data_series
Events /events events.get
KPIs /kpi/new_users/data_series kpi.new_users.data_series
KPIs /kpi/dau/data_series kpi.dau.data_series
KPIs /kpi/mau/data_series kpi.mau.data_series
KPIs /kpi/uninstalls/data_series kpi.uninstalls.data_series
Messages /messages/scheduled_broadcasts messages.schedule_broadcasts
Media Library (write) /media_library/create media_library.create
Preference Center /preference_center/v1/list preference_center.list
Preference Center /preference_center/v1/{preferenceCenterExternalID} preference_center.get
Purchases /purchases/product_list purchases.product_list
Purchases /purchases/revenue_series purchases.revenue_series
Purchases /purchases/quantity_series purchases.quantity_series
Segments /segments/list segments.list
Segments /segments/data_series segments.data_series
Segments /segments/details segments.details
Sends /sends/data_series sends.data_series
Sessions /sessions/data_series sessions.data_series
SDK Authentication Keys /app_group/sdk_authentication/keys sdk_authentication.keys
Subscription /subscription/status/get subscription.status.get
Subscription /subscription/user/status subscription.groups.get
Templates /templates/email/list templates.email.list
Templates /templates/email/info templates.email.info
Templates (write) /templates/email/create templates.email.create
Templates (write) /templates/email/update templates.email.update

[!WARNING] Do not reuse an existing API key—create one specifically for your MCP client. Assign least-privilege, non-PII permissions only. Write-capable scopes (rows marked "(write)" above: media_library.create, templates.email.create, templates.email.update, content_blocks.create, content_blocks.update) let agents create or modify content in your Braze workspace — only grant the specific write scopes you intend the agent to exercise, and consider a sandbox workspace for testing them.

[!NOTE] For /media_library/create, the server supports either asset_url or base64-encoded file uploads (asset_file_base64). Uploaded media must follow Braze media library constraints (for example image type and size limits, including the 5 MB image limit documented by Braze).

Step 2: Install uv

Run the following command to install uv, a command-line tool by Astral for dependency management and Python package handling:

MacOS and Linux

curl -LsSf https://astral.sh/uv/install.sh | sh

Windows

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Step 3: Configure your MCP client

Configure your MCP client using our pre-provided configuration file.

Claude

  1. In Claude Desktop, go to Settings > Developer > Edit Config, then add the following snippet:
    {
      "mcpServers": {
        "braze": {
          "command": "uvx",
          "args": ["--native-tls", "braze-mcp-server@latest"],
          "env": {
            "BRAZE_API_KEY": "your-braze-api-key",
            "BRAZE_BASE_URL": "your-braze-endpoint-url"
          }
        }
      }
    }
    
  2. Save the configuration and restart Claude Desktop.
  3. To verify your connection, try asking a question like "List my Braze campaigns".

Cursor

  1. In Cursor, go to Settings > Tools and Integrations > MCP Tools > Add Custom MCP, then add the following snippet:
    {
      "mcpServers": {
        "braze": {
          "command": "uvx",
          "args": ["--native-tls", "braze-mcp-server@latest"],
          "env": {
            "BRAZE_API_KEY": "your-braze-api-key",
            "BRAZE_BASE_URL": "your-braze-endpoint-url"
          }
        }
      }
    }
    
  2. When you're finished, save the configuration and restart Cursor.
  3. To verify your connection, try using the provided MCP tools to interact with your Braze data.

Build MCP Bundle (.mcpb)

You can package this server as an MCP Bundle for easy installation in compatible MCP clients.

Prerequisites

  • Node.js and npm installed (for MCPB CLI)
  • uv installed and on PATH (required at runtime by the server)
    • macOS/Linux:
      curl -LsSf https://astral.sh/uv/install.sh | sh
      
    • Ensure ~/.local/bin is on your PATH

Install MCPB CLI

npm install -g @anthropic-ai/mcpb

Package the bundle

From the repository root:

# Create source directory for packaging
mkdir -p dist/mcpb-source
cp -r src/ dist/mcpb-source/
cp pyproject.toml dist/mcpb-source/
cp uv.lock dist/mcpb-source/
cp manifest.json dist/mcpb-source/
cp README.md dist/mcpb-source/

# Generate the MCPB bundle
mcpb pack dist/mcpb-source/ dist/braze-mcp-server-source.mcpb

This creates a file at:

./dist/braze-mcp-server-source.mcpb

Install in MCP Clients

Option 1: Claude Desktop (Recommended)

  1. Open the .mcpb file directly with Claude Desktop
  2. Claude will automatically handle installation and prompt for configuration
  3. When prompted, enter:
    • BRAZE_API_KEY: your Braze API key (read‑only permissions only)
    • BRAZE_BASE_URL: your Braze REST endpoint (e.g., https://rest.iad-01.braze.com)
  4. Restart Claude Desktop if the server doesn't start automatically

Option 2: Other MCP Clients (Cursor, etc.)

  1. Unpack the bundle:
    mcpb unpack dist/braze-mcp-server-source.mcpb /path/to/install
    
  2. Configure your MCP client to use the unpacked server
  3. Set environment variables in your client configuration

Troubleshooting

MCPB CLI Issues

  • "mcpb: command not found":
    • Install MCPB CLI: npm install -g @anthropic-ai/mcpb
    • Verify installation: mcpb --version

Server Runtime Issues

  • "spawn uvx ENOENT" or "uvx not found":
    • Install uv using the command above
    • Make sure ~/.local/bin is on PATH

Bundle Installation Issues

  • "Manifest validation failed":
    • Check that manifest.json is valid JSON
    • Ensure all required fields are present in the manifest
  • "Cannot open .mcpb file":
    • Verify your MCP client supports MCPB bundles
    • Try manual installation with mcpb unpack

API Connection Issues

  • "Invalid API key" or "Authentication failed":
    • Verify your BRAZE_API_KEY is correct and has read permissions
    • Check that BRAZE_BASE_URL matches your Braze instance
  • "Connection timeout":
    • Verify your Braze instance URL is correct
    • Check network connectivity and firewall settings

Release flow

See the Release Process section in DEVELOPMENT.md for the canonical steps, including the two version-bump locations (pyproject.toml and manifest.json) and how the Test PyPI / PyPI workflows are triggered.

Disclaimer

The Model Context Protocol (MCP) is a newly introduced open-source protocol that may be susceptible to security issues or vulnerabilities at this time.

Braze MCP Server setup code and instructions are provided by Braze “as is” and without any warranties, and customers use it at their own risk. Braze shall not be responsible for any consequences arising from improper setup, misuse of the MCP, or any potential security issues that may arise. Braze strongly encourages customers to review their configurations carefully and to follow the outlined guidelines to reduce risks associated with the integrity and security of their Braze environment.

For assistance or clarification, please contact support@braze.com.

Metadata

Release files for braze-mcp-server 1.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for braze-mcp-server 1.2.0
File Size Uploaded
braze_mcp_server-1.2.0.tar.gz 322.4 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for braze-mcp-server 1.2.0
File Interpreter ABI Platform
braze_mcp_server-1.2.0-py3-none-any.whl Python 3 none any Details

Total release size: 383.8 kB

Release files / braze_mcp_server-1.2.0.tar.gz

Download URL braze_mcp_server-1.2.0.tar.gz
Size 322.4 kB
Tags Source
SHA-256 checksum
How to use checksums
faeea1b8f60fed19403bd0fc04e2c9912716c9b79c88262fab810f659e675938
BLAKE2b-256 checksum
How to use checksums
cc4cb9a4826b015a47647ae326f59119f79a5dc6d5666a7c9e5e69598bc12b2a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.13.12

Release files / braze_mcp_server-1.2.0-py3-none-any.whl

Download URL braze_mcp_server-1.2.0-py3-none-any.whl
Size 61.4 kB
Tags Python 3
SHA-256 checksum
How to use checksums
908639021905df008eb90fa6f752f4ad07341e42beb250bad97037f0ca552836
BLAKE2b-256 checksum
How to use checksums
6e578ab9862f3aab777ce6f4b6bea372ceb7652c7e7671c4a56821e3607fba3f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.13.12

Release history Release notifications | RSS feed

This release

1.2.0 This release

2 release files

1.1.0

2 release files

1.0.4

2 release files

1.0.1

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page