Embrace MCP Server is now GA

Customers are already querying the Embrace MCP Server to access data on over 70 apps, using AI clients like Claude, Cursor, OpenCode, Windsurf, Codex, and more! It launched with the following six tools:

• `list_apps`: Find and search applications in your Embrace workspace
• `get_app_details`: Get health metrics, crash-free rates, and session counts for an app
• `get_top_versions`: Identify which app versions are most widely used
• `list_crashes`: List top crashes ranked by frequency and user impact
• `get_crash_details`: Get detailed information about a specific crash group
• `get_crash_stack_samples`: Fetch actual stack traces for crash analysis

While we’ve seen regular usage from all our available tools, the most popular tools are the ones for investigating crashes.

See how this works in practice below.

This demo shows connecting to the Embrace MCP Server, viewing available tools, and using natural language to explore what data you can access.

A common workflow here is starting with a health check to look for any regressions. If you notice a crash spike, you can prioritize new ones by user impact and then start your investigation by pulling stack traces directly into your terminal. If you need more context about the app, device, or user experience to track down the root cause, you can then head to the Embrace dashboard to review full User Timelines.

What’s new in the Embrace MCP Server

Crashes are just the beginning. We’ve now added network performance tools to the Embrace MCP Server so you can quickly check network endpoint health, review trends, and break down errors and latency across multiple dimensions like country, OS, device, app version, and more. 

Here’s the list of new network performance tools you can use today: 

• `list_network_domains`: List all domains the app communicates with
• `list_network_endpoints`: List endpoints ranked by latency/errors/volume
• `get_network_endpoint_errors`: Get error breakdown (4xx, 5xx, connection)
• `get_network_endpoint_latency`: Get latency distribution and percentiles
• `get_network_endpoint_breakdown`: Break down by country, OS, device, etc.
• `get_network_endpoint_timeseries`: Get metrics over time       

Here are just a few questions you can answer with this data:

• Which endpoints have the highest latency, and how are they trending over time?
• Which endpoints have the highest error rate, and what’s the breakdown of these different error types?
• Are network failures disproportionately affecting certain types of users?

Here’s what this looks like in practice.

In this demo, a natural language question triggers MCP to analyze network endpoints and return ranked issues, error rates, latency percentiles, and first- vs third-party domain insights.

Check out these new network performance tools and share your feedback with us! Let us know what you find most useful and where we can improve them. 

What’s coming up next in the Embrace MCP Server 

Embrace excels in our collection and enrichment of high-fidelity production telemetry that is grounded in real user behavior. Our goal is to make as much of this data available for advanced analysis as possible so that you can surface real insights and prioritize problems based on how they’re affecting your users.

In other words, we’ve only scratched the surface on what’s possible with the Embrace MCP Server. 

This quarter, we are working on more robust support for web applications as well as creating tools for all the other event types you’re familiar with (e.g., spans, logs, sessions, and more). As the name suggests, context is key, and our goal is to help boost your AI client’s context with richer data throughout. This will be more than just raw data, it will include derived insights that further maximize your agent’s ability to clearly understand, interpret, and take action based on the information we provide through the MCP server.

If you’d like to try out the Embrace MCP Server, head to our docs to learn how to configure it, explore supported use cases, and start integrating Embrace observability data into your AI workflows.