Skip to content

HiwiKInsight 1.0.0

Choose a tag to compare

@github-actions github-actions released this 09 Oct 09:07
· 14 commits to main since this release

First open-source release of HiwiKInsight: self-hosted product analytics and App Store revenue for indie iOS apps, built to be queried by AI agents as well as people.

Try it

docker run --rm -p 8080:8080 ghcr.io/redhiwik/hiwikinsight:v1.0.0 demo -data /data -listen :8080

Then open http://localhost:8080/dashboard/ and sign in with demo / demo. Or download an archive below (it contains hiwikinsight, insight and config.example.yaml), or install the CLI with go install github.com/RedHiwiK/HiwiKInsight/cmd/insight@v1.0.0.

Swift SDK: HiwiKInsightKit 0.2.0. Guides: Getting started, Deployment, AI integration.

What's included

  • Single Go binary (hiwikinsight) with embedded pure-Go SQLite (no cgo), subcommands serve, demo, check-config, hash-password and version.
  • One YAML configuration file with ${NAME} / ${NAME:-default} environment expansion, strict validation, configurable reporting time zone, language (en/zh) and base currency.
  • Event ingestion (POST /v1/events) for the HiwiKInsightKit Swift SDK (protocol schema 1): raw DEFLATE, limits, deduplication, clock correction, per-install daily cap, no IP addresses stored.
  • Derived tables for installs, daily activity and purchase attempts; analysis views v_user_days, v_user_summary, v_sessions, v_screen_usage; raw event retention (retention_days).
  • "Truly new user" detection from the App Store original download date.
  • Environment safety nets: production data from never-released builds and from development devices is reclassified as sandbox.
  • App Store Server Notifications V2 (POST /v1/appstore/notifications) with JWS verification, idempotent storage, base-currency conversion, transaction emails and purchase attribution through appAccountToken.
  • App Store Connect sync: Summary Sales Reports (daily and monthly history), Analytics Reports (impressions, product page views, first-time downloads by source and country) and the released-build list.
  • 15 built-in metrics with fixed definitions: active_users, activity_days, retention, returning, module_usage, funnel, revenue, distribution, overview, sessions_daily, users, event_trend, errors, portfolio, store_funnel.
  • Read-only query API (/v1/query/describe, meta, metric/{name}, sql, user/{id}) with bearer tokens, a read-only SQLite connection and limits (single SELECT/WITH, 1000 rows, 10 s, 400-day range).
  • describe semantic layer for AI assistants: conventions, tables, views, metrics, data volume, undocumented events and per-app event catalogs.
  • insight CLI and a built-in MCP server (insight mcp) with the tools describe, list_apps, metric, sql and user.
  • Claude Code skill hiwikinsight-analytics.
  • Web dashboard with 13 pages (Portfolio, Overview, Engagement, Retention, Modules, Users, User detail, Acquisition, Revenue, Devices & Regions, Events, Errors, Settings), English and Chinese, light and dark mode, display currencies, bcrypt sign-in.
  • Daily and weekly email reports with preview and resend endpoints.
  • Alerts for failed purchases, payment errors, new and spiking errors, stale App Store Connect reports and silent apps, deduplicated per day.
  • Demo mode with three sample apps and 90 days of generated data (hiwikinsight demo, docker-compose.demo.yml, make demo).
  • Deployment assets: multi-arch container image on GHCR, Docker Compose with Caddy, systemd unit, nginx example, release archives for linux and darwin (amd64, arm64).
  • Documentation in docs/, AGENTS.md for AI coding agents, English and Chinese READMEs.