Capture the why behind AI-generated code. Generate decision logs from Claude Code sessions for PR reviews.

Reusable agent rules, templates, subagents, and skills for AI-assisted software delivery.

An append-only ledger that records why decisions were made.

Self-hosted database and API for AI project memory. Store decisions with receipts, query from any model

An append-only reasoning log for LLM-assisted research. Record the why behind decisions — with falsifiability conditions, confidence tags, and failure-first philosophy.

Capture decisions with full context, metadata, and confidence scores. Includes outcome tracking, notes, and CSV export.

Session Continuity Protocal (SCP): a lightweight protocol to prevent multi-session AI “amnesia” by persisting project state (PMP), decisions (ADR-lite), and no-silent-changes guardrails for consistent, token-efficient continuity.

Decision-first experiment tracker for product teams (hypothesis → decision → learning).

decisions, retros, and templates you can reuse.

AI assumption firewall for MCP agents — intercepts risky decisions, checks past decisions, detects contradictions, and gates human approval via Telegram or Slack.

Open specification for tamper-evident AI decision logs. Satisfies EU AI Act Articles 12 and 19. Defines record fields, hash chaining, Ed25519 signatures, and audit trail export format.

Minimal, zero-dependency decision log for Claude Code. Markdown-only, git-trackable, three smart hooks.

Your code knows what. Lore knows why. — CLI tool that captures the why behind code changes.

Persistent decision memory for any project. Karpathy's LLM Wiki, productized.

A platform that makes delivery friction visible, measurable, and actionable — so teams can fix real root causes instead of guessing what went wrong.

Git-native decision memory for AI coding agents.