Developer docs.

Install

One command. Auto-detects your IDE, builds a starter factbook from your repo, wires the MCP server. Runs entirely on your machine — no signup, no cloud account, no telemetry.

# macOS · Linux · WSL
curl -fsSL https://kernora.ai/install | sh

First run

Run nora init inside any repo. Nora scans your code, ADRs, tickets, and docs, builds a starter factbook in .nora/, and wires the MCP server to your installed editors.

cd ~/code/payments-api
nora init
→ Scanning repo for ADRs, tickets, conventions...
→ Found 18 candidate facts. Promoting 12 to canonical.
→ Detected editors: Cursor, VS Code, Claude Code.
→ MCP server registered. Restart your editor to activate.
✓ Factbook ready at .nora/factbook.jsonld

Editor integrations

Kernora ships with native integrations for the most common AI-assisted dev environments. Each integration is a thin MCP client — the factbook itself is identical.

Editor	Integration	Status
Claude Code (CLI)	--mcp-config .nora/mcp.json	Live
Cursor	Settings → MCP Servers → add Kernora	Live
VS Code (Copilot)	Extension marketplace ↗	Live · v2.7.0
Kiro	Built-in (v0.4+)	Live
JetBrains	Plugin marketplace	Beta
Zed	Native MCP	In development

Entry shape

Every fact in the factbook is a JSON-LD record. Plain markdown plus a structured index — sitting in .nora/ next to your code.

{
  "@id": "f234",
  "@type": "f:fact",
  "statement": "Stripe webhook handler must be idempotent via the event id.",
  "decided_on": "2026-03-12",
  "source": "docs/adr/018-webhook-idempotency.md",
  "supersedes": ["f009"],
  "applied": 42,
  "accept_rate": 0.94,
  "confidence": "canonical",
  "privacy": "team"
}

Trust ladder

Facts enter the factbook as proposed. The session analyzer promotes them after N applications without contradiction. You approve canonical from the dashboard.

State	Promotion rule	Visible to AI?
proposed	Just extracted — awaiting evidence	No
candidate	Applied ≥3× without contradiction	Yes, with candidate badge
canonical	Human approval from dashboard	Yes
superseded	Replaced by newer fact, history preserved	No (chain visible)

Emit formats

Same factbook. Five surfaces. Pick the formats you need — Kernora regenerates them on every change.

nora emit --format claude-code
nora emit --format cursor-rules
nora emit --format kiro-steering
nora emit --format copilot-instructions
nora emit --format claude-auto-memory  # NEW · v0.6

Inspector — verify after it writes

Grounding gets the right facts into the model. Inspector checks what comes back out: it reads the agent's output against the factlets that apply and returns a verdict — so "looks right" becomes "provably conforms, or here's exactly what broke."

The evaluation engine (conformance_verifier.py) is deterministic wherever a factlet carries a verify block. Dispatch is by verify.kind:

verify.kind	How it decides
assertion	A pure expression evaluated against the artifact — no eval/exec, just a sandboxed AST walk.
examples	given → expect cases; an uncovered boundary returns NEEDS_REVIEW, never a pass.
test_ref	Runs a referenced test through a caller-provided runner; flake or timeout → NEEDS_REVIEW.

Every run resolves to a closed verdict — PASS, FAIL, NEEDS_REVIEW, or NOT_APPLICABLE — and a FAIL carries a machine-readable rejection code plus the factlet it broke.

No synthetic passes. Where no deterministic oracle covers the artifact, Inspector returns NEEDS_REVIEW — it never fabricates a PASS from a model guess. Honesty over a green checkmark.

One engine, three cadences — a policy parameter, not three codebases: conformance (per change), compliance (evidence export), sweep (scan a whole repo).

Run it today: the live demo at trace.kernora.ai/inspector runs the engine on a real payload in your browser. An installed CLI gate, CI integration, and scheduled Sweep are on the roadmap.

Orchestrator runs the whole loop — ground → act → verify — with as much or as little human in it as your verification coverage earns. Today that means human-approves / human-reviews gates; fully unattended operation is on the roadmap.

Factlet protocol

The factbook is built on the open Factlet protocol. Five elements — f:fact, f:lineage, f:consent, f:confidence, f:emit — wired together with W3C PROV-O for lineage.

Your factbook is portable: an open spec, JSON-LD on disk, vendor-neutral. Switch tools, switch employers, switch models — the factbook moves with you.

MCP tools

Kernora exposes 50+ MCP tools. The core read surface works on every tier; mutation tools require Pro for team audit logs.

Tool	Purpose	Tier
nora_lookup	Query facts by topic, tag, or path	Free
nora_cite	Resolve a citation ID to its source + lineage	Free
nora_propose	Add a candidate fact from session output	Free
nora_supersede	Replace a canonical fact (audit-logged)	Pro
nora_consent	Set egress tier on a fact	Pro

nora_provenance

Return full PROV-O lineage for any citation. New in v0.6 — the tool that powers the inline citation chip in Nora Pro.

"name": "nora_provenance",
"input": { "fact_id": "f234" },
"output": {
  "statement": "...",
  "lineage": [
    { "role": "derivedFrom", "source": "docs/adr/018.md" },
    { "role": "supersedes",  "target": "f009" },
    { "role": "approvedBy", "agent": "alex@acme.dev", "at": "2026-03-22" }
  ],
  "confidence": 0.92
}

Team sync (Pro)

Pro adds encrypted S3 sync via the hosted relay at mcp.kernora.ai. Bring your own bucket; Kernora never holds the key. Per-fact egress consent is the default — you approve which facts cross the wire.

Security

BYOK end to end. On-device factbook. Per-fact egress consent. Audit log of every escalation. Built on the open Factlet protocol so you can verify the wire format yourself.

The install script is one line of sh — read it before you pipe it. The MCP server source is on GitHub. Run tcpdump on kernora.ai after install; on Free you'll see zero packets.

Telemetry

None on Free. On Nora Pro and Nora Team, telemetry is opt-in per category — crash reports, usage counts, model-routing stats — and listed in the dashboard with a one-click revoke. Nothing is sent without your consent; nothing identifies you by default.

Uninstall

One command:

kernora uninstall

This stops the daemon, removes ~/.kernora/, and unregisters the LaunchAgent.

Your .nora/ folder stays in your git repo — that's your data, you keep it. Re-running kernora install later picks up where you left off, with a fresh ~/.kernora/ cache.