Kernora — AI Work Intelligence
Kernora runs silently alongside your AI coding tools and turns every session into compounding intelligence. Patterns, decisions, and bugs are extracted into a local database and injected back into your agent's context automatically. Every session makes the next one smarter.
Your AI Leverage Score — a composite metric of prompt quality, context injection hit rate, decision acceptance, and pattern accumulation — starts at 1.0x and compounds toward 5.0x as Kernora learns your codebase.
No cloud. No API key required on Mac. Zero bytes sent to Kernora servers.
Install
VS Code / Kiro / Cursor:
Install from the Marketplace, or install the .vsix directly:
Extensions → Install from VSIX → select kernora-*.vsix
Kernora bootstraps automatically — creates a Python venv, installs deps, starts the dashboard at localhost:2742.
Claude Code:
curl -fsSL https://raw.githubusercontent.com/kernora-ai/nora/main/install.sh | bash
First run:
nora scan ~/code/your-project
Seeds your database from git history so Kernora has context from session one.
Dashboard
Open http://localhost:2742 to see:
| Tab |
What it shows |
| Home |
AI Leverage Score, loop health, top projects, rule suggestions |
| Projects |
Per-project AI metrics, patterns, decisions, bugs |
| Activity |
Session history with outcome indicators |
| Coach |
AI Leverage sparkline, coaching notes, certificate export |
| Knowledge |
Best practices, playbooks, anti-patterns |
| Memory |
Context injection feed, steering file viewer |
| Decisions |
Searchable architectural decisions |
| Bugs |
Bug inventory with severity, fix suggestions, mark resolved |
| Settings |
LLM provider config, local AI status |
AI Leverage Score
AI Leverage = 1.0 + (composite_quality × 4.0)
composite_quality = (prompt_quality × 0.4)
+ (injection_hit_rate × 0.3)
+ (decision_acceptance_rate × 0.2)
+ (pattern_accumulation_rate × 0.1)
| Score |
Label |
What it means |
| 1.0–2.0 |
Early |
AI isn't helping much yet |
| 2.0–3.0 |
Developing |
Getting value, room to grow |
| 3.0–4.0 |
Strong |
Measurably effective AI usage |
| 4.0–5.0 |
Excellent |
Elite AI collaboration |
Export your score as a shareable certificate from the Coach tab.
What You Can Say to Nora
All 18 tools are available as natural-language commands in your IDE's AI chat.
Explore Your History
| Command |
What It Does |
nora stats |
Session count, token usage, model breakdown over time |
nora search <query> |
Full-text search across patterns, decisions, bugs |
nora session <id> |
Full detail on a specific session |
Learn From Your Codebase
| Command |
What It Does |
nora patterns |
Recurring engineering patterns from your sessions |
nora decisions |
Architectural decisions with rationale |
nora bugs |
Past bugs with fix suggestions and severity |
nora skills |
Distilled team methodology — engineering rules and playbooks |
nora scan <path> |
Import a git repo's history (run once per project) |
Quality & Reviews
| Command |
What It Does |
nora pe-review <focus> |
Principal Engineer 4-tier code audit |
nora coe <issue> |
Blameless root-cause investigation (5 Whys) |
nora coe product <issue> |
Product COE — why a feature shipped wrong |
nora retro |
Engineering retrospective with git velocity metrics |
nora scope <task> |
Validate a task against project history before starting |
Factory & Coaching
| Command |
What It Does |
nora sofac |
Software Factory health — what shipped, what's pending (GREEN/YELLOW/RED) |
nora inventory |
Feature audit: SHIP/POLISH/WIRE/BLOCKER |
nora coach |
AI Leverage coaching — patterns, anti-patterns, before/after examples |
nora onboard |
Onboard a new developer with your team's methodology |
Help
| Command |
What It Does |
nora help |
Full tool reference with examples |
LLM Provider Priority
Nora tries these in order — the first available one wins:
- IDE LLM (VS Code, Kiro, Cursor) — zero config
- Apple FoundationModels (macOS 26+) — on-device, zero cost
- MLX-LM (macOS 14+) — on-device, ~2GB one-time download
- BYOK — Anthropic, OpenAI, Google, Bedrock, Grok
- Ollama — local, free
On a modern Mac, Kernora works with no API key.
Privacy
All data stays in ~/.kernora/echo.db on your machine. Zero bytes reach Kernora servers in BYOK mode. Analysis uses your own API key — the same call you'd make directly.
Architecture
- Database:
~/.kernora/echo.db (SQLite, WAL mode)
- Dashboard: Flask + HTMX at
localhost:2742
- MCP server: 18 tools via stdio JSON-RPC
- Hooks: 6 Claude Code hooks, 5 Kiro hooks
- Steering: Auto-generated markdown files injected into AI context
- Config:
~/.kernora/config.toml
Links
