All signal. No feed.

4DA reads the internet for developers — privately, locally — and gets sharper every day.

It scans your codebase — Cargo.toml, package.json, go.mod, Git history — and scores every article, advisory, and release from 20+ sources against what you actually build. An item needs 2+ independent signals to survive. Everything else is rejected.

Typical rejection rate: 99%+. What's left is yours.

Privacy-first. Local-first. BYOK. Your indexed content stays on your machine — there is no 4DA-operated server for it to go to. It learns from how you engage with what it shows you — yesterday's noise becomes tomorrow's signal. Crash reporting is opt-in, off by default. For the full list of outbound connections with source-code references, see NETWORK.md.

Try it in 30 seconds

Already using Claude Code, Cursor, or Windsurf? One command:

npx @4da/mcp-server --setup

This auto-detects your editor, scans your project, and gives your AI assistant context-aware developer intelligence. No desktop app required — works standalone. Learn more about the MCP server.

How It Scores

5 independent signal axes. An item must pass 2 or more to surface. Single-axis matches are hard-capped at 28% — no matter how strong one signal is, it cannot pass alone.

What passes the gate goes through 12 quality multipliers: content depth, novelty detection, competing tech penalties, title-body coherence, and intent scoring from recent work. Every constant is calibrated across 9 simulated developer personas with 215 labeled test items.

LLM Verification (Optional)

After keyword scoring, an optional LLM layer verifies the top items against your full developer context — stack, dependencies, recent commits, anti-technologies, and engagement history. Strict 1-5 rubric:

• 5 = MUST-READ: Security alert for YOUR dependency, breaking change YOU must act on
• 3 = WORTH KNOWING: Useful tool that fits YOUR exact stack
• 1 = NOISE: Mentions your tech but isn't actionable

This is where gold nuggets surface — articles the keyword pipeline misses because there's no keyword overlap, but the LLM understands the conceptual relevance to your specific project.

You control the compute. Use Ollama for free local inference (fully private), or bring your own Anthropic/OpenAI key. 4DA never pays for your compute, never stores your keys remotely, never makes API calls you didn't configure. No LLM = keyword pipeline only (85-90% accurate).

It Gets Smarter

Every interaction is a training signal:

• Save → topics boost, source reputation rises, taste embedding sharpens
• Dismiss → topics suppress, anti-patterns form, future noise drops
• Autophagy cycles detect systematic over/under-scoring and self-correct
• Topic decay calibrates per-topic: "Rust" stays relevant for days, "JS framework of the week" decays in hours

... View full README on GitHub