Discover which Keypup solution fits your workflow best. Compare our interactive web platform with AI-powered dashboards against our MCP Server for seamless AI assistant integration. Both leverage the same powerful analytics engine.
TL;DR: Keypup offers two powerful ways to access your engineering analyticsâeach with distinct strengths.
Engineering leaders face a common challenge: accessing the right data at the right time to make informed decisions. Keypup solves this through a powerful analytics engine that harmonizes data across GitHub, GitLab, Jira, Azure DevOps, and more. But we offer two distinct ways to interact with this engineâeach optimized for different workflows.
This isn't about choosing between "good" and "better." It's about understanding which tool fits your specific needs. Some scenarios demand the interactive, visual-first experience of our web platform. Others benefit from the seamless AI integration our MCP Server provides.
Let's explore both approaches, understand their unique strengths, and help you decide which (or both) will transform your engineering analytics workflow.
Before diving into the differences, it's crucial to understand what makes both approaches powerfulâthey're built on the same foundation:
The Challenge: Your engineering data is scattered across multiple toolsâGitHub, Jira, GitLab, Azure DevOps, Bitbucket, and more. Each has its own data model, terminology, and quirks.
The Solution: Keypup's data harmonization engine automatically:
The Result: A unified view of your SDLC, regardless of which tools you use.
Both the platform and MCP Server leverage the same sophisticated calculation engine that powers:
Whether you're using the web platform's AI agent or querying through the MCP Server, both support natural language queries:
"Show me our PR cycle time trend over the last quarter, broken down by team"
No SQL. No complex query builders. Just ask.
The Keypup web platform is your command center for engineering analytics. It's designed for interactive exploration, visual discovery, and real-time monitoring.
â AI-Powered Dashboard Generation: Describe what you want to see, and our AI agent builds the perfect dashboard in seconds
â Interactive Drill-Down: Click any data point to explore the underlying detailsâsee the actual PRs, commits, or issues
â Live, Dynamic Data: Dashboards refresh automatically. Always see your current metrics without manual updates
â Visual-First Experience: Rich, interactive charts with hover details, filters, and cross-dashboard navigation
â Collaborative Workspaces: Share dashboards with your team, set up alerts, and collaborate on insights
â Historical Analysis: Time-travel through your data with date range selectors and comparison views
The web platform is ideal when you need to:
Let's see the platform handle a real-world scenario.
Prompt to Keypup Platform AI Agent
"Create a dashboard showing our PR review process health for Q2 2026. I need to see review turnaround time, reviewer load distribution, and identify any bottlenecks. Break it down by team."
What Happened on the Platform
Instant Dashboard Creation:
Interactive Capabilities:
Live Data:
Key Insights Surfaced:
Follow-Up Prompt
"Show me the 12 PRs waiting for review over 48 hours. I want to see their size, author, and when they were created."
Drill-Down Details
What the Platform Provided:
Immediate Actions Enabled:
This type of drill-down from summary to detail is where the platform excels.
The MCP Server takes a fundamentally different approach: instead of bringing you to the data, it brings the data to wherever you're already workingâyour AI assistant of choice.
â AI-Native Integration: Query Keypup data directly from Claude, ChatGPT, Cursor, Windsurf, or any MCP-compatible AI tool
â Context Mixing: Combine Keypup metrics with external dataâmarket trends, customer feedback, or other business metrics
â Custom Visualizations: AI generates charts tailored to your exact reporting needs, not pre-built templates
â Workflow Integration: Get insights without context-switchingâstay in your IDE, terminal, or chat interface
â Flexible Output: Generate markdown reports, HTML visualizations, CSV exports, or inline summaries
â Scriptable Analysis: Programmatically query Keypup in your own automation workflows
The MCP Server is ideal when you need to:
Let's use the same scenario but through the MCP Server in Claude Desktop.
Prompt to Claude Desktop (with Keypup MCP Server)
"Using the Keypup MCP server, show me our PR review process health for Q2 2026. I need review turnaround time, reviewer load distribution, and bottlenecks by team. Create an HTML visualization."
What Happened with MCP Server
AI-Generated Analysis:
Key Differences from Platform:
query_dataset, list_dataset_fieldsWhat the AI Provided:
Unique Capability: The AI can now combine this with other context. For example: "Compare these review metrics with our deployment frequency data and explain if there's a correlation."
Here's where the MCP Server truly differentiatesâcombining multiple data sources:
Advanced MCP Prompt
"I've attached our Q2 customer satisfaction scores (CSV). Using Keypup MCP data, analyze if there's a correlation between our PR review turnaround time and customer satisfaction. Create a report with visualizations."
Multi-Source Analysis
What the AI Did:
This type of cross-domain analysis is impossible with the platform alone.
The MCP Server lets you bring engineering metrics into conversations about business outcomes, customer impact, or strategic planningâall within your AI assistant.
| Capability | Keypup Platform | MCP Server |
|---|---|---|
| Data Currency | Live, real-time data with auto-refresh | Snapshot at query time (not live) |
| Visual Exploration | Interactive dashboards with drill-down | AI-generated static visualizations |
| Context Switching | Requires opening web browser | Works in your existing AI tool |
| Data Mixing | Keypup data only | Mix with external data sources |
| Collaboration | Share live dashboards with teams | Share generated reports/visualizations |
| Drill-Down | Click to explore individual records | Request details via follow-up prompts |
| Monitoring | Always-on dashboards for team use | On-demand queries |
| Automation | Scheduled reports via platform | Scriptable via MCP protocol |
| Learning Curve | Web UI + dashboard concepts | Natural language prompts only |
| Output Formats | Platform dashboards | HTML, Markdown, CSV, images |
The answer isn't "one or the other"âmost teams benefit from both. Here's how to think about it:
đ You need ongoing monitoring: Set up dashboards for daily standups, weekly reviews, or executive updates
đ You're investigating issues: Drill into anomalies, explore patterns, and discover root causes interactively
đĽ You're collaborating with teams: Share live dashboards that everyone can reference and explore
đ You track improvement initiatives: Monitor KPIs over time with always-current data
đź You present to leadership: Show live dashboards in meetings where stakeholders can click and explore
đ¤ You work primarily in AI tools: Stay in Claude, ChatGPT, or your IDE without context-switching
đ You create custom reports: Build executive summaries combining narrative and data
đ You mix data sources: Correlate engineering metrics with business outcomes or customer data
⥠You need quick answers: Get specific metrics without loading a full dashboard
đ ď¸ You automate workflows: Integrate Keypup data into your own scripts or tools
Most engineering leaders use both:
Sarah, the engineering manager, opens the Keypup platform on her second monitor. Her "Daily Team Health" dashboard loads automatically:
She shares her screen during stand-up. The team discusses the 3 delayed PRs and assigns reviewers on the spot.
During a Slack conversation about slower-than-expected feature delivery, Sarah asks Claude:
"Using Keypup MCP, compare our PR cycle time this month vs. last month, broken down by feature vs. bugfix PRs"
Claude responds in 10 seconds with:
Sarah shares this analysis directly in Slack. No dashboard needed.
Intrigued by the PR size finding, Sarah opens the platform's "PR Size Analysis" dashboard. She:
She schedules a 1-on-1 with that team lead to discuss breaking down work.
Sarah needs to send a weekly update to her VP. She prompts Claude:
"Using Keypup MCP, create a markdown report for the week of June 10-16, 2026. Include:
Format it for our weekly executive email."
Claude generates a polished markdown report in 15 seconds, including narrative context. Sarah makes minor edits and sends it.
{
"mcpServers": {
"keypup": {
"url": "https://hq.keypup.io/mcp",
"headers": {
"Authorization": "Bearer YOUR_KEYPUP_API_TOKEN"
}
}
}
}
Absolutely. They're complementary, not competing. Many teams:
Yes. Both query the identical data harmonization engine. A metric queried via MCP Server will match the same metric on a platform dashboard (accounting for query time differences).
Both are included in your Keypup subscription. No additional cost for using one vs. the other.
Yes. The visualizations and reports generated via MCP can be saved as HTML, images, or PDFs and shared with anyone. They don't need AI assistant access.
The platform dashboards load quickly because they're optimized for visual display. MCP queries take 5-15 seconds because the AI is interpreting your request, querying data, and generating custom output. Both are fast enough for real-time decision-making.
Keypup's dual approach gives you flexibility that no other engineering analytics platform offers:
Both are powered by the same world-class data harmonization and calculation engine. Both support natural language queries. Both deliver actionable insights in seconds, not hours.
The question isn't which to chooseâit's how to leverage both for maximum impact.
Ready to transform your engineering analytics workflow? Start with the platform for your core dashboards, then add MCP Server integration to supercharge your AI assistant.
Get started today: https://hq.keypup.io/signup
Platform: Data refreshes every 15 minutes at least, or immediately for source tools with webhooks. Dashboards update automatically in your browser.
MCP Server: Queries always fetch the latest data at query time (no caching).
Both platform and MCP Server support:
The Keypup MCP Server implements the full Model Context Protocol specification and works with:
Want to see Keypup in action? Book a demo or start your free trial today.
Join teams already using AI to make data-driven decisions faster than ever.
Discover how Keypup's Model Context Protocol (MCP) server lets you query your engineering data from Claude, ChatGPT, Cursor, and any MCP-compatible AI client using natural language. No code, no SQL, no complexity.
Discover how Keypup's MCP Server integration brings engineering analytics directly into your AI workflow with natural language queries, generating sophisticated visualizations in seconds. Explore 5 real-world use cases from cycle time analysis to comprehensive SPACE metrics.
Unlock efficiency by mastering your pull request workflow. Learn how Keypup's AI unifies Git and Jira data to track, analyze, and optimize your entire issue-to-merge lifecycle.