R0LL0/research-radar

Research Radar is a local-first research intelligence app: it ingests RSS and arXiv feeds, stores items and extracted text in SQLite, and generates AI-written briefings where each claim links to a verbatim source quote. It’s aimed at individual operators or teams who need a private command center for tracking technical research without losing provenance. Key features: - Ingest: RSS + arXiv, deduplication, article text extraction, source health - Briefings: one-click briefs, batch “brief all new”, HN+article “pulse” - Precise citations: click a quoted sentence to jump and highlight the source - Intelligence: signal scores, story clustering, watchlist terms, daily digest, webhook - Discovery/export: find feeds from a site, export Markdown or print-friendly HTML - Automation: auto-ingest interval, cron/CLI hook (/api/cron/ingest) - Auth/workspaces: per-user workspace isolation, email/password, session cookies - AI: default local Ollama or BYOK cloud keys (OpenAI/Anthropic/etc.) Stack: Next.js + React + TypeScript, Prisma + SQLite (single-user), scrypt passwords and AES-256-GCM for stored keys. Deployment notes: SQLite for local use; switch to Postgres for multi-tenant hosting. Requirements: Node 20+, Ollama recommended.

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.