Product analytics you can just ask.
Funnels, retention, paths, and cohorts from one snippet. Ask "where do trials drop off?" in plain english, from a dashboard bar or your own Cursor or Claude over MCP. Answers are computed from the reports, never guessed, with a verdict on what to fix. One Go binary, cheap, self-host free.
What is product analytics?
Product analytics is the practice of measuring how people actually use a product by tracking the actions they take inside it, then turning those events into funnels, retention, paths, and cohorts. Where web analytics counts pageviews and traffic sources to tell you how people arrive, product analytics counts named events (signed up, activated, upgraded) keyed to a stable per-user id to tell you what they do once inside, and whether they stick.
The two are most useful together: a user's referrer, their first click, and their eventual upgrade belong on one timeline, not in two tools that never agree. smolanalytics does both from one snippet. The full definition, and exactly how it differs from web analytics, is on the product analytics glossary page.
What are the core reports?
Almost every product question resolves to one of four reports. Learn these and you can read a product's health without a data team.
All four run off the same raw material: named events with a stable distinct_id. Send that well once and every report becomes a question you can just ask. See the whole set on every feature.
How smolanalytics does product analytics
The report list is not the difference, every tool has funnels and retention. What is different is how you get the answer, and what it costs:
- 1Ask in plain English. Instead of building a dashboard for every question, you type "where do trials drop off?" into a dashboard bar, or into your own Cursor or Claude Code over MCP (47 tools, 13 prompts). It is bring-your-own-model, so your editor's own AI does the asking and the AI part is free.
- 2A verdict, not just charts. Beyond the reports, it tells you what to fix, on the dashboard and in a morning brief. The point of product analytics is a decision, so it surfaces the decision instead of another graph to interpret.
- 3Computed, never guessed. Every answer comes from the same deterministic reports the dashboard renders, not from an LLM writing numbers. A CI agreement test fails the build if the AI answer ever differs from the dashboard, so the number you get is the real one.
- 4One binary, and cheap. A single MIT Go binary, standard library only, roughly 7 bytes per event, no Kafka, ClickHouse, or Postgres to run. Plans start at $9/mo with $5 per extra million events (the big tools charge around $50), and self-host is free forever.
It deliberately does not do session replay, feature flags, experiments, heatmaps, or surveys. Cookieless mode means no consent banner, and importers for PostHog, Umami, CSV, and JSONL get your history in. See how it stacks up vs Mixpanel and vs Amplitude.
Ask it like you'd ask a data person
The questions you actually have never fit a pre-built chart. So type them:
Same reports either way. The dashboard ask bar is the fastest path; the MCP server answers the same questions in the window where you write code. The full connect walkthrough is on the MCP server page.
Run the whole thing with one line
Web install is a single script tag plus one init call. The dashboard, ingestion, and the MCP server are all in one binary, no infra to stand up:
<script src="https://smolanalytics.com/sdk.js"></script>
<script>smolanalytics.init("YOUR_KEY", { host: "https://your-instance" })</script>docker run -p 8080:8080 ghcr.io/arjun0606/smolanalytics # dashboard on http://localhost:8080, ingestion at /v1/events, MCP server ready to connect
Prefer real data first? The live demo is a populated instance you can ask right now, no install. Full setup is in the docs, and self-host notes are in the GitHub README.