iroha

Sessions scatter. iroha remembers.

English | 日本語

Persist Claude Code sessions to Notion as a living, queryable team memory — decisions (with rationale and rejected alternatives), work-state, chat-style highlights, and per-project architecture profiles. So humans and future Claude sessions can recall what was decided, why, what's unfinished, and how things are built.

Why

Claude Code's built-in memory is thin and lives on one machine. iroha turns each coding session into structured, searchable memory in Notion, so that the next session — yours or a teammate's — opens already knowing the project's decisions, its unfinished work, and how it is built. The more the team uses it, the more it grows: ask "have we built something like this before?" and iroha points at the prior session, the files it changed, and why.

How it works

• The runtime is Bun + TypeScript (scripts/**/*.ts, run directly with no build step):

it does the deterministic extraction (changed files, commands, metadata) from the session transcript. The intelligence (summary, decisions, classification, chat highlights) is produced by Claude inside the skills.

• All Notion reads/writes go through the Notion MCP — there is no API token.

Auth is the MCP's OAuth, so setup is a single connection. Works on the free Notion plan.

• Recall is two-stage. On every prompt, a UserPromptSubmit hook runs a cheap, local

BM25 search (search.ts, CJK-aware, no LLM and no network) over a tiny on-disk index and proactively surfaces the most relevant past decisions — so Claude consults them before rebuilding, at zero per-prompt latency or token cost. When that pointer isn't enough, /iroha:recall escalates to Notion semantic search (notion-search, free plan) for the full rationale and rejected alternatives. The free tier carries most of the weight; the semantic stage catches the paraphrases it misses. Optional: a local hybrid tier (armed once by the bun "<plugin>/scripts/rerank-setup.ts" command /iroha:init prints — runs from any project, installs into the plugin not your cwd) — a dense bi-encoder (multilingual-e5-small) generates the semantic near-matches BM25 can't (zero lexical overlap), and a cross-encoder reranker (bge-reranker-v2-m3) promotes the strong matches above the BM25 list. Measured on this repo's index: it recovers a candidate-generation MISS the lexical stage can't (Recall@3 86%→93%) while keeping cross-domain abstention at 100%. The reranker only promotes, never vetoes a BM25 hit — vetoing cost real recall (a terse true match scores like an off-topic one). Opt-in and heavy (two local models, ~700MB total) — a fresh install stays on the dependency-free BM25 tier and pays nothing.

• A SessionStart hook injects the project's State (from a small repo mirror) so

Claude proactively tells you where you left off and what's unfinished. After /compact or auto-compact it also re-injects the current session's own thread (your prompts + a capped recent tail), so the conversation survives compaction.

• A write-time check (a PreToolUse hook on git commit) runs the same cheap local

recall over the commit subject + staged paths and, if Active decisions govern that area, adds an advisory to verify you are not silently reversing one (and to run /iroha:check). It is purely advisory — it never blocks or auto-approves the commit — and catches a course-reversal at the last moment before code lands.

• Each saved session also carries a metrics dashboard (turns, tool calls, files,

duration) and a collapsed full-chat audit trail. /iroha:digest rolls a week or month into one page; /iroha:audit keeps the growing memory clean (duplicate decisions, State drift, stale items).

Memory model — three layers + State

graph TD
  CC["Claude Code session"] -->|/iroha:save-session| SK["save-session skill"]
  SK -->|deterministic| EX["extract.ts (Bun)"]
  SK -->|intelligence| CL["Claude"]
  SK -->|Notion MCP / OAuth| N[("Notion")]
  N --> SES["Sessions — what happened"]
  N --> DEC["Decisions — why"]
  N --> PRJ["Projects — current stack"]
  N --> ST["State — where we are"]
  ST -->|repo .iroha/state.md| HK["SessionStart hook"]
  DEC -->|notion-search| RC["/iroha:recall"]

• Sessions — what happened each session: summary, decisions made, chat highlights, changed files.
• Decisions — why the project is the way it is: rationale + rejected alternatives, with supersession history (a change of mind is itself memory).
• Projects — what the project is now: languages, key libraries, dev tooling, CI, an architecture diagram — for onboarding and cross-project search.
• State page — the always-current "where are we / what's unfinished", injected at session start.

Requirements

• Claude Code
• A Notion account with the hosted Notion MCP connected (OAuth). Works on the free plan.

Install

In Claude Code:

/plugin marketplace add hir4ta/iroha-for-session
/plugin install iroha@iroha-for-session

Getting started

1. Connect Notion MCP — run /mcp, pick notion, and complete the OAuth in your browser.
2. `/iroha:init` — creates the Sessions / Decisions / Projects databases (plus

Recent / Active / By-Language views) under a Notion page you choose. Re-running it on a shared page lets a teammate join the same workspace.

1. `/iroha:save-session` — save the current session.
2. `/iroha:recall <query>` — "did we decide against X? why?" / "have we built this before?".
3. `/iroha:project` — record (or refresh) the project's tech stack. Manual, engineer-reviewed.

Commands

What iroha is not

• No secrets. No API token to manage — Notion auth is MCP OAuth only; only non-secret

ids are cached locally.

• No relation properties. Session↔Decision links use a URL property (a known

relation-write bug in the Notion MCP); promotable to native relations once stable.

• No verbatim transcript dump. The curated highlights are the headline — their

You lines anchored to your real messages (never invented). A cleaned, per-turn-capped full-chat audit trail sits collapsed underneath, with thinking and tool noise stripped.

• No save coercion. Hooks remind, they don't block.

Design

• Architecture invariants: .claude/rules/architecture.md
• Project notes & scope: CLAUDE.md
• Contributing: CONTRIBUTING.md · Security: SECURITY.md

License

MIT © Shunichi Hirata