4DA Signal

The intelligence layer
your tools can't replicate

4DA already scores content against your codebase across 10 sources. Signal adds the analysis engine that connects the dots, detects your blind spots, and gets smarter with every interaction.

// what signal does in a sentence
input10 sources × your codebase context × time
output: connected intelligence, blind spot detection, domain evolution tracking
deltacompounds daily — week 1 is useful, month 3 is indispensable

Most developers never need more than this.

The free tier is a complete product. AI briefings, scoring, context discovery, and the full STREETS Playbook are included forever. No trial. No account. No expiration.

AI daily briefings
AI weekly digest
10 content sources
PASIFA scoring engine
ACE context scanner
STREETS playbook (7 modules)
Behavior learning
MCP server (31 tools)
Feed UI + saved items
CLI binary
Runs 100% locally
Unlimited use

The analysis engine that turns data into decisions.

Every feature below works locally, uses your own API keys, and produces intelligence that is impossible to get any other way — because it requires your codebase context and your interaction history to exist.

Signal Chains

See the full story, not fragments

A Tokio release drops. Three hours later, a blog post explains the migration path. By evening, two HN threads debate the trade-offs and a Reddit thread surfaces a workaround for a breaking change. You see all four — separately. Signal Chains connects them into one linked narrative, ranked by relevance to your Cargo.toml. One click, complete picture.

chains = cross-source entity linking + temporal clustering + codebase affinity scoring
Knowledge Gaps

What you don't know — quantified

Your package.json says you use Next.js 14. You've shipped App Router pages. But Knowledge Gaps notices you've never engaged with any content about Server Actions, Partial Prerendering, or the new caching semantics. Those aren't random omissions — they're blind spots in features you're actively shipping against. Knowledge Gaps surfaces them before they become production incidents.

gaps = dependency graph analysis × content engagement history × topic coverage scoring
Developer DNA

Your technical identity, auto-generated

Developer DNA builds a living profile of your expertise: 68% systems (Rust, WASM, networking), 24% web (TypeScript, React), 8% infrastructure (Docker, CI/CD). It tracks depth, not just breadth — you're not just a "Rust developer," you're deep in async runtimes and shallow on macros. Watch it evolve as your work evolves. Export as SVG card for your README, markdown for your portfolio, or keep it private.

dna = manifest analysis + interaction patterns + expertise depth weighting | export: svg, md, card
Semantic Shifts

Watch your domain evolve in real time

Three months ago, "edge runtime" appeared in 2% of your feed. Now it's 19%. Semantic Shifts pinpoints the inflection to a specific Vercel blog post on March 3rd that triggered a conversation cascade across 4 sources. Trend detection that's personal to your stack, not the internet's.

shifts = embedding trajectory analysis over sliding time windows per topic cluster
Natural Language Search

Ask your content history anything

"What was that article about SQLite WAL mode performance from last week?" Signal searches across all 10 sources, your saves, your dismissals, and your interaction history. Semantic search, not keyword matching. It finds the article even if you never saved it — because you read it for 4 minutes.

search = vector similarity + interaction-weighted ranking + cross-source semantic retrieval
Project Health

Early warning for your dependencies

Your Cargo.toml has 47 dependencies. Project Health monitors their pulse: commit velocity, issue trends, release cadence, maintainer activity. When a critical dep goes quiet for 3 weeks after 6 months of weekly releases, you know before the GitHub issue asking "is this project dead?" appears.

health = dependency manifest parsing × upstream signal monitoring × anomaly detection
Score Autopsy

Full transparency on every score

An item scored 0.87. Why? Score Autopsy shows the full breakdown: Relevance 0.92 (3 matching technologies in your manifest), Freshness 0.85 (posted 4 hours ago), Authority 0.78 (author maintains a dep you use), Proximity 0.91 (direct stack overlap), Action Potential 0.89 (migration guide format). Every axis, every weight, every signal. No black boxes.

autopsy = full PASIFA 5-axis decomposition with per-signal attribution and weight explanation
Standing Queries

Persistent watchlists that never sleep

Set "Tauri security advisories" as a standing query. Every new item from every source gets tested against it. When CVE-2026-XXXX drops at 3am and appears on a security blog, Standing Queries flags it before your next coffee. Like Google Alerts but running against scored, codebase-relevant content.

standing = persistent semantic queries × real-time new-item evaluation × notification triggers
Attention Report

See where your focus actually goes

You spent 71% of your reading time on Rust content this month, but your project is 55% TypeScript. Your attention doesn't match your work distribution. Attention Report shows the disconnect — and whether it's a problem (neglecting your primary stack) or a signal (you're preparing to migrate).

attention = interaction time tracking × stack distribution analysis × attention/work divergence scoring
Decision Signals

Content that arrives at the right moment

You've been reading 6 articles about Hono vs Express this week. Decision Signals detects the pattern and proactively surfaces the 3 most authoritative comparisons, a benchmark from last month you missed, and a migration guide from a team that just shipped the switch. The right information at the decision point.

decisions = reading pattern clustering + decision-point detection + proactive content surfacing

Signal gets smarter. Every other tool stays static.

Every save, every dismiss, every click, every minute spent reading teaches Signal about your work. The intelligence compounds. Here's what that looks like over time:

Day 1 Codebase scan

ACE reads your manifests. Signal knows your stack, your dependencies, your language distribution. Scoring begins immediately. You see relevant content within 60 seconds.

Week 1 Behavior learning kicks in

Signal learns what you engage with and what you skip. Developer DNA begins forming. Knowledge Gaps start identifying areas where your dependency graph exceeds your reading history.

Month 1 Patterns emerge

Signal Chains connect items across sources. Semantic Shifts detect topic trajectories. Decision Signals recognize when you're evaluating a technology choice. Attention Reports show where your focus diverges from your work.

Month 3 Compounding intelligence

90 days of interaction data. Signal anticipates your needs. It surfaces content you didn't know you needed. Knowledge Gaps close. Your Developer DNA shows expertise trajectory. Standing Queries have caught things you would have missed. The tool that was useful on day 1 is now indispensable.

Year 1 A second technical brain

365 days of codebase-aware, privacy-preserving intelligence. Semantic Shifts show you how your entire domain evolved. Developer DNA charts your growth. Signal Chains have connected hundreds of narratives. No other tool on Earth has this data, because no other tool runs locally against your codebase for a year.


Your intelligence stays yours. Always.

Signal runs 100% locally. Your codebase, your interaction history, your Developer DNA, your Knowledge Gaps — none of it touches our servers. The license key is the only network call Signal makes to us.

AI features use your own API keys (BYOK). Anthropic, OpenAI, or Ollama for fully local inference. We never see your prompts. We never see the responses. We are architecturally unable to access your data.

Cancel your subscription and your data stays. Your interaction history, your Developer DNA, your entire intelligence graph — all of it remains on your machine. We unlock the analysis engine. The data is always yours.


For developers who want to know how.
01 / Ingestion

10 sources, unified pipeline

Hacker News, Reddit, GitHub, RSS, Lobsters, DevTo, and more. Each source adapter normalizes content into a common schema. Fetched on your machine, stored in SQLite locally.

02 / Scoring

PASIFA 5-axis engine

Every item scored on Proximity, Action Potential, Significance, Information Freshness, and Authority. Weighted by your interaction history. Multi-signal confirmation gate rejects noise. 99%+ of content is filtered out.

03 / Intelligence

Signal analysis layer

Cross-source entity linking builds Signal Chains. Embedding trajectory analysis detects Semantic Shifts. Dependency graph diffing surfaces Knowledge Gaps. All computed locally using sqlite-vec for vector operations.


One tier. No gotchas.

The free tier is a complete product. Signal is for developers who want compound intelligence.

$12/mo
 
$12 / 160 work hours = $0.075/hr for an intelligence system that watches your ecosystem 24/7
45-day free trial included in the app. No credit card required to try.
  • Everything in the free tier
  • Developer DNA profiling + exports
  • Signal chain analysis
  • Knowledge gap detection
  • Semantic shift tracking
  • Natural language search
  • Standing queries (persistent watchlists)
  • Score autopsy (full 5-axis breakdown)
  • Project health monitoring
  • Decision signals
  • Attention analytics
  • Priority support

All prices in AUD. Cancel anytime. Your data never leaves your machine.


Questions
What's the difference between free and Signal?
Free gives you the full scoring engine, AI briefings, all 10 sources, and the complete STREETS Playbook. Signal adds the intelligence analysis layer: Signal Chains, Knowledge Gaps, Developer DNA, Semantic Shifts, Natural Language Search, Project Health, Standing Queries, Score Autopsy, Attention Reports, and Decision Signals. Free surfaces relevant content. Signal tells you what it means, what you're missing, and how your domain is evolving.
Wait, AI briefings are free?
Yes. Daily and weekly AI-generated briefings are included in the free tier. They use your own API key (BYOK) or local Ollama. We believe every developer deserves an AI-generated summary of what matters to their stack. Signal adds the intelligence that makes those briefings richer — connected chains, blind spot detection, and trend context.
What happens to my data?
Nothing. 4DA runs 100% locally. Signal features use your own API keys (BYOK) to call AI models. Raw data never leaves your machine. The license key is the only thing that touches our servers.
Do I need an API key?
The scoring engine and content surfacing work without any API key. AI-powered features (briefings, natural language search, score autopsy) need an LLM provider: OpenAI, Anthropic, or Ollama for fully local inference. You provide your own key in settings. We never see or store it.
Why does Signal cost money if it uses my API key?
Your API key powers the models. Signal provides the intelligence architecture built on top — the cross-source entity linking, the dependency graph analysis, the temporal embedding tracking, the interaction-weighted ranking algorithms. Think of it as: you bring the engine, Signal is the dashboard, the instruments, and the navigation system.
Can I try before subscribing?
Yes. 4DA includes a 45-day free trial of all Signal features. No credit card required. Install the app, enable Signal features, and evaluate with your real codebase and real content. The compound advantage means the trial gets more valuable every day you use it.
What if I cancel?
Your license remains valid until the end of the billing period. After that, you revert to the free tier. No data is lost — your interaction history, your Developer DNA, your entire intelligence graph stays on your machine. You just lose access to the analysis tools that query it.
How is this different from GitHub Copilot / Cursor / etc?
Those tools help you write code. 4DA helps you know what to build and why. It operates on the information layer: what's happening in your ecosystem, what you're missing, how your domain is evolving. They're complementary. 4DA even ships an MCP server with 31 tools so your AI coding assistant can draw from your scored intelligence feed.
Is there a team plan?
Team and Enterprise tiers are coming. If you need multi-seat licensing now, get in touch.