Talivia: Self-Hosted Analytics That Ties Traffic to Revenue

1 min read
self-hostingrevenue-attributionstripemcp-server
View as Markdown
Originally from github.com
View source

My notes

Summary

Talivia is an open-source, self-hosted analytics platform positioned as a “revenue-first” alternative to Datafast. It bundles web analytics, session replay, and customer revenue attribution (Stripe, LemonSqueezy, Polar, Dodo, Yolfi, or a manual payment API) in one self-hosted app, with a paid cloud tier for extra integrations such as Google Search Console, Bing Webmaster Tools, GitHub activity, and social mentions.

Key Insight

  • Ties raw traffic analytics directly to revenue: first-touch and last-touch attribution, subscription lifecycle, refunds, and disputes are tracked alongside standard web analytics and session replay. Most self-hosted analytics tools (Plausible, Umami) stop at pageviews and do not connect to payment providers.
  • Deployment is genuinely lightweight: Node.js 22/24 LTS, pnpm 10+, one PostgreSQL database, and one required secret (APP_SECRET, 32-byte random). The Docker Compose path applies migrations automatically on container start.
  • Ships an “Agent Kit” MCP server (npx -y @talivia/agent mcp or the hosted OAuth endpoint https://talivia.com/mcp) built for Claude Code, Codex, ChatGPT, and other MCP clients. It installs tracking snippets, generates framework-specific setup plans, and verifies live analytics events are firing. That is a template worth noting: shipping an agent-facing setup and verification tool alongside the product itself, not just human docs.
  • The default bootstrap login is admin/admin, a real deployment risk if the “change it immediately” step gets skipped, worth flagging for anyone spinning this up unattended.
  • No migration path exists from a hosted Talivia database to self-hosted, and migrations must never be hand-edited once applied. Standard, but easy to violate under time pressure.