competitor-launch-radar

What problem does it solve?

Competitor launch radar identifies newly launched agent-frameworks that are not part of the already-tracked cohort, so you can notice fresh entrants as soon as they appear on Product Hunt or Hacker News.

Core Features & Use Cases

• Weekly scans across two launch surfaces: Pulls Product Hunt RSS and Hacker News Algolia results, limited to the last 7 days, to find new frameworks.
• Cohort suppression to avoid noise: Filters out known frameworks in a hardcoded “tracked cohort” list so only true new entrants are surfaced.
• Deduplicated, one-time announcements: Uses a local LRU state file to ensure each entrant is reported exactly once (no re-announce churn).
• Classification for operator clarity: Categorizes entrants as framework, mcp, or product using keyword + heuristic priority (framework > mcp > product).
• Safety-oriented, untrusted-input handling: Treats feed/API text as untrusted and includes explicit rules to omit flagged “instructions-like” content.

Quick Start

Run competitor-launch-radar in normal mode (with the default empty var) to generate this week’s digest and notifications for any new entrants found.