Open Source Project IntelligenceAll projectsRead as Markdown

Read the evidence behind an open source project.

An evolving platform that turns repository activity into evidence about project health, momentum, sustainability and abandonment risk.

Public demo not configured.Source on GitHub

The product

Open Source Project Intelligence collects repository activity and turns commits, issues, pull requests, releases and contributor signals into evidence about health, sustainability, momentum and abandonment risk. Public catalog and project details coexist with authenticated workspace and administration tools.

  • Go
  • React
  • TypeScript
  • PostgreSQL
  • NATS JetStream
  • S3
  • Valkey
  • OpenTelemetry
Public demo not configured.Source on GitHub

The main flow

  1. Collect and normalize

    A bounded worker pipeline collects GitHub activity through an adapter and converts provider data into canonical records, preserving collection evidence.

  2. Calculate deterministic signals

    Versioned definitions, time windows and cutoffs make metrics reproducible. Unknown, not applicable and insufficient data are distinct states; missing data never becomes zero.

  3. Compare and explain

    Project detail, comparisons and governed AI runs expose evidence. Human approval bounds assistant actions; exports preserve durable results.

Architecture

GitHub → collection adapter → bounded workers → canonical PostgreSQL records

PostgreSQL outbox → NATS JetStream · evidence → S3 · disposable cache → Valkey

React ↔ versioned Go API → metrics / comparisons / governed AI runs

Open Source Project Intelligence

A modular Go core

Separate API and worker binaries share business packages. PostgreSQL owns durable data; a transactional outbox delivers through NATS JetStream. S3 stores evidence bytes and Valkey provides disposable acceleration.

Evidence before interpretation

React uses a generated versioned HTTP client. Metrics and synchronization decisions are deterministic; model unavailability affects AI analysis rather than those calculations.

Availability and limits

An evolving implementation, not a guarantee about a project’s future. Metric availability depends on source evidence and collection windows. Workspace and administration require authentication. A public demo is shown only when its destination is configured.

Public demo not configured.Source on GitHub