Business judgment on AI products
Independent sellers see which page brought a paid order, and can just ask what happened yesterday.
01
Start inside the user's day · Public facts + observable behavior · 2026-09-01
Independent sellers need to see which page brought a paid order, and can ask in natural language what happened yesterday.
Currently may use privacy-friendly analytics tools like Plausible, but lack revenue attribution and natural language querying.
Traditional analytics tools struggle to attribute traffic to specific revenue, and querying data requires complex report operations.
The trend is web analytics leaving pageview dashboards for “which page made money,” queryable by AI. Don't polish another traffic chart. Start with indie stores, courses, and plugin checkout. Self-host is free; hosted starts at nine dollars a month.
Its public repository has 371 stars, showing developer attention; repeat use and payment are not yet verified.
① Whether GitHub stars keep rising in three months — is the MCP interface really pulling developers; ② Whether anyone publicly says "migrated from Plausible to getopen," and why; ③ How accurate attribution is on real order data — can it explain refunds and subscription discounts
Worth trying. Its public repository has 371 stars, showing developer attention; repeat use and payment are not yet verified.
a "deliberate blanks" list of the incumbent — enumerate what the existing tool skips because of its own positioning, then pick two blanks as your selling points. Any AI product positioned as "an X alternative" can copy this analysis frame directly. free self-host plus hosted event tiers ($9/$19/$29) and a one-year free Startup plan. Standard and clear, worth referencing.
Free to self-host (AGPL-3.0); hosted from $9/month (the author's own words on HN: "self-host it for free, or $9 hosted"). Third-party listings show Hobby $9 (100k events/mo), Starter $19 (500k), Growth $29 (1M), plus a free Startup plan (1 … ① Whether GitHub stars keep rising in three months — is the MCP interface really pulling developers; ② Whether anyone publicly says "migrated from Plausible to getopen," and why; ③ How accurate attribution is on real order data — can it explain refunds and subscription discounts
Independent sellers see which page brought a paid order, and can just ask what happened yesterday.
Its public repository has 371 stars, showing developer attention; repeat use and payment are not yet verified.
An English validation note will follow from the public evidence.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
02
Market comparison
Local supply: Emerging
Demand evidence: Not yet verified
Public coverage has been recorded for this market. · 2026-09-01
03
The call and next move come first; the full read retains the evidence and counterevidence.
A modern alternative to Plausible: cookieless open-source web analytics with two things nobody else ships — revenue attribution and an MCP server, so you can ask Claude what happened yesterday and get the number back.
The OpenLabs team (HN handle rahulbridge). GitHub org OpenLabs-so; the main repo is openanalytics (TypeScript, AGPL-3.0), plus a macOS menu-bar client, oa-menubar. The repo was created 2026-08-11 and reached 117 stars in three days.
Read: the tool itself is not new; the two additions are what matter for the AI era. Dashboards are table stakes; revenue attribution and an MCP interface are the differentiation it is actually betting on.
What it deliberately does not do: no personal data collection, no cross-site cross-day profiles. The page's line: "privacy is the default, not a setting."
The GA-era workflow: drop in a heavy GA4 script → get forced into a cookie consent banner → wait two days → wander a maze of a dashboard to find "which page brings the most signups." Plausible already killed half of that (no cookies, light script), but it stops at the pageview layer.
What getopen replaces is the second half: knowing "which page brings the most paid signups" used to mean manually joining order exports, or adopting heavyweight analytics like Amplitude. Revenue attribution compresses that into one dashboard dimension; the MCP step goes further and turns "looking at data" from manual labor into a conversation.
Free to self-host (AGPL-3.0); hosted from $9/month (the author's own words on HN: "self-host it for free, or $9 hosted"). Third-party listings show Hobby $9 (100k events/mo), Starter $19 (500k), Growth $29 (1M), plus a free Startup plan (1 year, 100 projects) — the official pricing page was not directly captured; treat actuals as canonical.
Read: $9 entry with event-based tiers is the standard Plausible/Umami playbook, no price war intended. The real moat bet is "revenue attribution + agent interface" — which neither Plausible nor Umami has.
| Dimension | Call |
|---|---|
| Founder-product fit | The author ships AI tools continuously (org name OpenLabs); knows developer distribution |
| Product insight | In a crowded analytics lane, picked exactly the two gaps Plausible leaves: revenue attribution and an agent-facing interface |
| Execution quality | ClickHouse backend plus MCP plus six-channel attribution; 117 stars in three days signals developer approval |
| Timing | Analytics is a red ocean, but "AI agents must be able to read your data" just became a requirement — good positioning |
It is betting on the next era's question: data is not just for humans to look at, it is for AI agents to read.
Analytics has churned through three rounds (GA → Plausible/Umami → PostHog), and the dashboards are all copies of each other. getopen's smart move is not a better dashboard; it is two interfaces no one else ships — revenue attribution wires analytics to money, and MCP wires analytics to agents. For an indie developer, "which page brought N orders yesterday" is the only question worth asking daily, and Plausible cannot answer it.
The transferable rule: in a red ocean, find the two things the previous generation deliberately does not do — do not try to do it better. Plausible deliberately skips revenue (privacy-first, simple) and skips agent interfaces (its user is a human). getopen turned both gaps into its selling points. Any "alternative to X" positioning should start by listing X's deliberate blank spots.
The risks: revenue attribution across six processors is data-dirty in production (event-order joins, refunds, subscription discounts) — a long war; and the HN comments name Umami's free tier as a direct competitor outside price. The moat is not yet proven.
① Whether GitHub stars keep rising in three months — is the MCP interface really pulling developers ② Whether anyone publicly says "migrated from Plausible to getopen," and why ③ How accurate attribution is on real order data — can it explain refunds and subscription discounts
Product logic: a "deliberate blanks" list of the incumbent — enumerate what the existing tool skips because of its own positioning, then pick two blanks as your selling points. Any AI product positioned as "an X alternative" can copy this analysis frame directly.
Pricing structure: free self-host plus hosted event tiers ($9/$19/$29) and a one-year free Startup plan. Standard and clear, worth referencing.
Worth watching. In an analytics lane full of identical dashboards, it carved out real differentiation with revenue attribution plus an agent interface, and 117 stars in three days is the market's first signal. Keep it as the case study for "how an alternative finds differentiation"; verify against the three checks in three months.
04
Evidence trail
05
Use these searches when the official site is missing or the current link is only a lead.