Heap MCP Integration

Weldable's Heap MCP integration connects your AI agents to your Heap product analytics data. Your agent can query user behavior, pull event data, check funnel metrics, and retrieve session information through natural language. Instead of building charts in the Heap UI every time someone has a product question, your agent gets the answer directly from the data.

Heap's signature feature is automatic data capture: every click, pageview, form submission, and interaction is recorded without manual instrumentation. This means the data your agent can query is far richer than what you'd get from a manually tagged analytics tool. Connecting Heap to Weldable through the MCP protocol turns that broad dataset into something your agent can search, filter, and report on across all your connected tools.

Use cases

Product usage monitoring

Your agent queries Heap for daily active users, feature adoption rates, and engagement metrics, then posts a morning summary to your product team's Slack channel. It compares today's numbers against the 7-day rolling average and flags any metric that dropped more than 10%. Product managers see the health of their features without building a custom dashboard.

Funnel analysis on demand

When someone asks "what's the conversion rate from signup to first purchase this month?", your agent queries Heap's funnel data with the appropriate event filters and date range. It returns the overall conversion rate, the drop-off at each step, and how the numbers compare to last month. Getting a funnel answer takes seconds instead of the 15 minutes it would take to set up the analysis in Heap's UI.

Feature launch tracking

After shipping a new feature, your agent monitors adoption by querying Heap for the specific interaction events tied to that feature. It tracks how many unique users triggered the event each day, segments by user cohort (new vs. existing), and reports the results daily for the first two weeks. The product team gets a clear picture of whether the feature is gaining traction without setting up a separate tracking plan.

User session investigation

When a customer reports an issue, your agent can pull their recent session data from Heap, including the sequence of pages visited, buttons clicked, and errors encountered. It formats this into a timeline that the support or engineering team can review. This cuts the time between "something's broken" and "here's what happened" from hours to minutes.

Retention cohort reporting

Your agent queries Heap for retention data, grouping users by their signup week and tracking how many return in subsequent weeks. It calculates week-1, week-4, and week-12 retention rates and compares them across cohorts. The results post to your growth team's channel, highlighting cohorts with notably better or worse retention so the team can investigate what drove the difference.

How it works

Connect your Heap account through Weldable's setup flow using your Heap API credentials. Heap's APIs use app-level authentication with an API key that corresponds to your project. Weldable stores the credentials securely and handles authentication on every request.

Once connected, describe what you need in plain English. Say "show me the top 10 most clicked buttons this week" or "how many users completed the onboarding flow yesterday?" and Weldable translates your intent into the appropriate Heap API calls. Results come back structured and ready for your agent to summarize, compare, or forward to another integration.

Tips

Heap captures everything, so be specific with your queries. Because Heap auto-captures all interactions, a broad query can return an overwhelming amount of data. Include date ranges, event types, and user filters in your requests to get focused results.

Heap's event taxonomy is based on interaction type. Events in Heap are categorized as pageviews, clicks, submits, and changes. When asking your agent about specific user actions, reference the interaction type to help it select the right event class. "How many users clicked the upgrade button" is more precise than "how many users interacted with the upgrade button."

Custom events and properties extend what your agent can query. While Heap captures interactions automatically, you can define custom events and user properties through the Heap API or SDK. These custom definitions show up in the same query interface, so your agent can filter by properties like plan tier or user role.

Heap Connect exports data to your warehouse. If you need your agent to run complex queries that go beyond the Heap API's capabilities, consider using Heap Connect to export data to a warehouse like BigQuery or Snowflake, then query it through a separate database connection.

Rate limits depend on your Heap plan. The API enforces rate limits that vary by plan tier. For high-frequency monitoring workflows, check your plan's limits and adjust your polling intervals accordingly. Batch your queries where possible to stay within the allowed request volume.