Show HN: Clamp – Web analytics your AI agent can read and query
A new web analytics tool called Clamp lets AI agents query your site's traffic data using plain English instead of dashboards. It connects to popular AI coding environments like Cursor and Claude Desktop through Anthropic's Model Context Protocol, making analytics a native part o...
According to the Clamp product page at clamp.sh, submitted to Hacker News by user sidneyottelohe, the platform is a fresh attempt to solve a problem that nobody talked about until AI agents made it obvious: most analytics tools are built for human eyes, not machine queries. Clamp flips that assumption entirely. It is built from the ground up as an MCP server, meaning AI assistants in Cursor, Claude Desktop, Windsurf, GitHub Copilot, and OpenAI's tooling can ask it questions and get structured answers back, no dashboard required.
Why This Matters
The standard analytics workflow, where a developer ships a feature and then separately logs into Google Analytics or Mixpanel to check the numbers, is already starting to feel antiquated. Clamp represents the next logical step: collapsing that loop so your AI coding assistant is also your analyst. With the MCP server ecosystem already tracking over 85,000 stars on the punkpeye awesome-mcp-servers GitHub repository, the protocol is clearly winning adoption fast. Any analytics vendor that does not offer a machine-readable interface in the next 12 to 18 months will look like the companies that built desktop software and ignored the browser.
Daily briefing from 50+ sources. Free, 5-minute read.
The Full Story
Clamp starts with a one-line npm install. You run npm i @clamp-sh/analytics, push your code, and events start flowing automatically. The platform auto-instruments page views, sessions, and referrers from the first deploy, which means you are not staring at empty charts waiting to configure tracking. For Next.js, React, or plain JavaScript projects, that zero-configuration promise is genuinely useful because setup friction is the reason most side projects never bother with analytics at all.
The core product experience lives inside your IDE. Instead of opening a browser tab to check how a new pricing page is performing, you ask your AI assistant directly. Clamp's demo shows a realistic example: a developer asks "how's /pricing doing?" and gets back a conversion rate that moved from 7.8 percent to 11.2 percent, a 43 percent lift, with a breakdown showing desktop drove the improvement while mobile was flat. That kind of segmented, context-aware answer would normally require you to build a custom report in a traditional analytics tool.
Funnel creation follows the same conversational pattern. You describe the steps in plain English, something like "signup, connect repo, first deploy," and Clamp maps out where users drop off at each stage. The example on the product page shows 2,847 users hitting signup, 64 percent making it to the repo connection step, and only 23 percent completing their first deploy. It also surfaces the median time between steps, which was 4 minutes and 12 seconds in the example shown.
The regression detection feature is where things get genuinely interesting for teams shipping fast. Clamp watches your key metrics automatically and surfaces alerts the next time you open your IDE, not as a notification that lands in Slack where it gets buried, but as part of your active conversation. The product demo shows Clamp identifying a 22 percent bounce rate jump on the signup page, narrowing it down to mobile Safari users with viewports smaller than 390 pixels, and linking the probable cause to a specific pull request, PR 214. It then offers to open a fix, which becomes PR 218 with a responsive CTA correction.
Privacy was clearly a design constraint, not an afterthought. Clamp is cookie-free, GDPR-compliant by default, hosted in the EU, and resolves IP addresses only down to the country level before discarding them. The tracking script itself comes in under 1 kilobyte, which means it will not move your Lighthouse performance score.
Key Details
- Free tier supports up to 5,000 events with no credit card required.
- Compatible with 5 AI platforms: Cursor, Claude Desktop, Windsurf, GitHub Copilot, and OpenAI tooling.
- Install requires a single npm package:
@clamp-sh/analytics. - The MCP server ecosystem reference repository has accumulated 85,000 GitHub stars and over 9,300 forks.
- The tracking script is sub-1 kilobyte, smaller than most favicon files.
- Custom event tracking uses a single API call:
clamp.track('signup', { plan: 'pro' }). - Hacker News submission received 2 points and 1 comment within approximately 15 hours of posting.
What's Next
Clamp will likely live or die by how deeply it integrates into developer workflows beyond the initial novelty of asking analytics questions in a chat window. The real test is whether teams adopt it for production monitoring, specifically the automated alerts and weekly summaries that show up in the IDE, which would make it sticky in a way that passive dashboards never achieve. Watch for competing moves from established players like Plausible and Fathom, both of which have already built privacy-first analytics but have not yet shipped MCP-native interfaces.
How This Compares
The obvious comparison is to Plausible Analytics and Fathom, the two most prominent privacy-first analytics tools that Clamp most resembles in terms of cookie-free tracking and GDPR positioning. Both are mature products with solid reputations, but neither has built an MCP layer. That gap is what Clamp is running through. Plausible does offer an API, but querying it requires custom code, not a natural language conversation inside your IDE.
PostHog is a more formidable comparison point. PostHog added AI features to its product analytics platform and supports custom event tracking at a level of depth that Clamp does not yet appear to match. But PostHog's product is complex, designed for product teams with dedicated analysts, not solo developers who want to ask a quick question without leaving their coding environment. Clamp is betting that simplicity and IDE-native access beat feature depth for a large slice of the market, and that bet is not unreasonable given how many developers never meaningfully engage with the analytics tools they install.
The broader MCP ecosystem context matters here too. Microsoft, Atlassian, Stripe, and dozens of other companies have shipped MCP integrations over the past several months, as covered in AI Agents Daily news. Analytics was an obvious gap in that ecosystem. Clamp is among the first to fill it with a product aimed at developers rather than enterprise buyers, which gives it a real shot at becoming the default choice before the larger platforms catch up. The AI tools landscape is moving fast enough that being first with a clean, working MCP integration genuinely matters right now.
FAQ
Q: What is Clamp and how does it work with AI agents? A: Clamp is a web analytics platform that functions as an MCP server, which means AI coding assistants like Cursor or Claude Desktop can query your site's traffic data using plain English. You install a single npm package, push your code, and your AI assistant can answer questions about page performance, conversion rates, and bounce rates directly inside your IDE.
Q: Is Clamp free to use for small projects? A: Yes. Clamp offers a free tier that supports up to 5,000 events per month with no credit card required. This makes it practical for side projects, early-stage startups, or developers who want to test the platform before committing to a paid plan.
Q: How is Clamp different from Google Analytics? A: Google Analytics is built around a visual dashboard designed for humans to click through and interpret. Clamp skips the dashboard entirely and instead lets AI agents query the data programmatically using natural language. It is also cookie-free and GDPR-compliant by default, which Google Analytics is not without significant additional configuration.
The direction Clamp is pointing toward, analytics as a service that AI agents consume rather than a dashboard humans navigate, will almost certainly define how the next generation of analytics products gets built. Check the AI Agents Daily guides for tutorials on setting up MCP servers and integrating them into your development workflow. Subscribe to the AI Agents Daily weekly newsletter for daily updates on AI agents, tools, and automation.
Get stories like this daily
Free briefing. Curated from 50+ sources. 5-minute read every morning.
