https://github.com/sdsrss/claude-mem-lite
Lightweight persistent memory system for Claude Code — FTS5 search, episode batching, error-triggered recall
https://github.com/sdsrss/claude-mem-lite
claude-code fts5 hooks mcp memory persistence sqlite
Last synced: 15 days ago
JSON representation
Lightweight persistent memory system for Claude Code — FTS5 search, episode batching, error-triggered recall
- Host: GitHub
- URL: https://github.com/sdsrss/claude-mem-lite
- Owner: sdsrss
- License: mit
- Created: 2026-02-08T07:10:37.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2026-05-23T20:11:33.000Z (25 days ago)
- Last Synced: 2026-05-23T20:25:52.784Z (25 days ago)
- Topics: claude-code, fts5, hooks, mcp, memory, persistence, sqlite
- Language: JavaScript
- Size: 3.11 MB
- Stars: 43
- Watchers: 0
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
Awesome Lists containing this project
README
[English](README.md) | [中文](README.zh-CN.md)
# claude-mem-lite
`claude-mem-lite` is a **persistent memory** (also called *long-term memory* or *cross-session context*) system for **[Claude Code](https://docs.anthropic.com/en/docs/claude-code)** — Anthropic's CLI coding agent. It runs as an **[MCP](https://modelcontextprotocol.io/) server** plus a set of Claude Code hooks, automatically capturing coding observations, decisions, and bug fixes during sessions, then providing hybrid full-text + semantic search to recall them later.
Compared to general-purpose LLM memory frameworks like [`mem0`](https://github.com/mem0ai/mem0) or the MCP reference [`memory`](https://github.com/modelcontextprotocol/servers/tree/main/src/memory) server, claude-mem-lite is purpose-built for Claude Code's hook lifecycle: episode batching cuts LLM calls 7–10× vs the original [claude-mem](https://github.com/thedotmack/claude-mem) (an estimated ~600× lower total cost — see the cost model below; this is an architecture estimate, not a measured benchmark), while the hybrid FTS5 + TF-IDF retriever benchmarks at 0.88 Recall@10 / 0.96 Precision@10.
> 中文简介：claude-mem-lite 是 Claude Code 的轻量级**持久化记忆 / 长期记忆 / 跨会话上下文**插件，基于 MCP 协议 + 钩子机制，自动捕获编码会话中的决策、修复和上下文，并通过 FTS5 + TF-IDF 混合检索召回。详见 [中文 README](README.zh-CN.md)。
Zero external services. Single SQLite database. Minimal overhead.
## Why claude-mem-lite?
A ground-up redesign of [claude-mem](https://github.com/thedotmack/claude-mem), replacing its heavyweight architecture with a smarter, leaner approach.
### Architecture comparison
| | claude-mem (original) | claude-mem-lite |
|---|---|---|
| **LLM calls** | Every tool use triggers a Sonnet call | Only on episode flush (5-10 ops batched) |
| **LLM input** | Raw `tool_input` + `tool_output` JSON | Pre-processed action summaries |
| **Conversation** | Multi-turn, accumulates full history | Stateless single-turn extraction |
| **Noise filtering** | LLM decides via "WHEN TO SKIP" prompt | Deterministic code-level Tier 1 filter |
| **Runtime** | Long-running worker process (1.8MB .cjs) | On-demand spawn, exits immediately |
| **Dependencies** | Bun + Python/uv + Chroma vector DB | Node.js only (3 npm packages) |
| **Source size** | ~2.3MB compiled bundles | ~50KB readable source |
| **Data directory** | `~/.claude-mem/` | `~/.claude-mem-lite/` (hidden, auto-migrates) |
### Token & cost efficiency
For a typical 50-tool-call session (illustrative cost model — the ratios below are
architecture estimates derived from batch size, token counts, and model pricing, **not**
a measured end-to-end benchmark):
| | claude-mem | claude-mem-lite | Ratio (estimated) |
|---|---|---|---|
| LLM calls | ~50 (every tool use) | ~5-8 (per episode) | **~7-10x fewer** |
| Tokens per call | 1,000-5,000 (raw JSON + history) | 200-500 (summaries only) | **~5-10x smaller** |
| Total tokens | ~100K-250K | ~1K-4K | **~50-100x less** |
| Model cost | Sonnet ($3/$15 per M) | Haiku ($0.25/$1.25 per M) | **~12x cheaper** |
| Combined savings | | | **~600x lower cost (estimated)** |
### Quality comparison
| Dimension | Winner | Why |
|---|---|---|
| **Classification accuracy** | Tie | Both produce correct type/title/narrative |
| **Noise filtering** | **lite** | Code-level filtering is deterministic; LLM "WHEN TO SKIP" is unreliable |
| **Observation coherence** | **lite** | Episode batching groups related edits into one coherent observation |
| **Code-level detail** | original | Sees full diffs, but rarely useful for memory search |
| **Search recall** | Tie | Users search semantic concepts ("auth bug"), not code lines |
| **Hook latency** | **lite** | Async background workers; original blocks 2-5s per hook |
### Design philosophy
The original sends **everything to the LLM and hopes it filters well**. claude-mem-lite **filters first with code, then sends only what matters** to a smaller model. This is not a downgrade; it's a smarter architecture that produces equivalent search quality at a fraction of the cost.
### Comparison: memory systems for AI coding agents
How claude-mem-lite differs from the major neighbors in the LLM-memory space (verified May 2026):
| | **claude-mem-lite** | [`mem0`](https://github.com/mem0ai/mem0) | MCP reference [`memory`](https://github.com/modelcontextprotocol/servers/tree/main/src/memory) | [claude-mem](https://github.com/thedotmack/claude-mem) (original) |
|---|---|---|---|---|
| **Target client** | Claude Code only | Any LLM app via SDK | Any MCP client | Claude Code only |
| **Capture model** | Auto via hooks | Manual `memory.add()` | Manual tool calls (`create_entities`, `add_observations`) | Auto via hooks |
| **Code-aware retrieval** | FTS5 + 100+ synonym pairs (incl. CJK↔EN) | General-purpose | Generic graph nodes | Code-aware |
| **Search** | Hybrid: FTS5 BM25 + TF-IDF cosine via RRF | Hybrid: semantic + BM25 + entity linking | Knowledge-graph traversal | FTS5 + Chroma vector |
| **Storage** | Single local SQLite | Pluggable; Qdrant or configurable vector store | Single JSONL file (knowledge graph) | SQLite + Chroma |
| **LLM dependency** | Haiku per episode (5–10 ops batched) | LLM per add/search op | None (graph CRUD only) | Sonnet per tool call |
| **Setup** | One command (`/plugin install` or `npx`) | SDK integration + vector store config | MCP install (per-client) | Bun + Python + Chroma |
**When to pick which**: pick `mem0` if you need a memory layer for a non-Claude-Code app (your own agent, multiple LLM providers). Pick the MCP reference `memory` server if you specifically want a knowledge-graph data model and don't mind invoking memory tools by hand. Pick claude-mem-lite if you want zero-touch automatic capture purpose-built for Claude Code's hook lifecycle, with code-domain retrieval and no external services.
## Features
- **Automatic capture** -- Hooks into Claude Code lifecycle (PostToolUse, SessionStart, Stop, UserPromptSubmit) to record observations without manual effort
- **Hybrid search** -- FTS5 BM25 + TF-IDF vector cosine similarity, merged via Reciprocal Rank Fusion (RRF). FTS5 handles keyword matching; 512-dim TF-IDF vectors capture semantic similarity for recall beyond exact terms
- **Timeline browsing** -- Navigate observations chronologically with anchor-based context windows
- **Episode batching** -- Groups related file operations into coherent episodes before LLM encoding
- **Error-triggered recall** -- Automatically searches memory when Bash errors occur, surfacing relevant past fixes
- **Proactive file history** -- When editing a file, automatically shows relevant past observations for that file
- **Session summaries** -- LLM-generated summaries at session end (via background workers using `claude -p`)
- **Project-scoped context** -- Injects recent memory into `CLAUDE.md` and session startup for immediate context
- **Observation types** -- Categorized as `decision`, `bugfix`, `feature`, `refactor`, `discovery`, or `change`
- **Importance grading** -- LLM assigns 1-3 importance levels (routine / notable / critical) to each observation
- **Observation relations** -- Bidirectional links between related observations based on file overlap
- **User prompt capture** -- Records user prompts via UserPromptSubmit hook for intent tracking
- **Read file tracking** -- Tracks files read during sessions for richer episode context
- **Zero data loss** -- If LLM fails, observations are saved with degraded (inferred) metadata instead of being discarded
- **Two-tier dedup** -- Jaccard similarity (5-minute window) + MinHash signatures (7-day cross-session window) prevent duplicates
- **Synonym expansion** -- Abbreviations like `K8s`, `DB`, `auth` automatically expand to full forms in FTS5 search (100+ pairs including CJK↔EN cross-language mappings)
- **CJK synonym extraction** -- Unsegmented Chinese text is scanned for known vocabulary words (数据库→database, 搜索→search, etc.) enabling cross-language memory recall
- **Stop-word filtering** -- English stop words filtered from both TF-IDF vocabulary (reclaiming ~18% of vector dimensions) and FTS queries (preventing false negatives from noise terms like "how", "the", "does")
- **Persisted vocabulary** -- TF-IDF vocabulary persisted to `vocab_state` table, preventing vector staleness when document frequencies shift. Vectors stay valid until explicit rebuild
- **Pseudo-relevance feedback (PRF)** -- Top results seed expansion queries for broader recall
- **Concept co-occurrence** -- Shared concepts across observations expand search to related topics
- **Context-aware re-ranking** -- Active file overlap boosts relevance (exact match + directory-level half-weight)
- **Superseded detection** -- Marks older observations as outdated when newer ones cover the same files with higher importance
- **Adaptive time windows** -- Session startup recall uses velocity-based time windows (high/medium/low activity tiers)
- **Token-budgeted context** -- Greedy knapsack algorithm selects session-start context within a 2,000-token budget, prioritizing by recency and importance
- **Observation compression** -- Old low-value observations can be compressed into weekly summaries to reduce noise
- **Secret scrubbing** -- Automatic redaction of API keys, tokens, PEM blocks, connection strings, and 15+ credential patterns
- **Atomic writes** -- All file writes (episodes, CLAUDE.md) use write-to-tmp + rename to prevent corruption on crash
- **Robust locking** -- PID-aware lock files with automatic stale/orphan cleanup (>30s timeout or dead PID)
- **Stale session cleanup** -- Sessions active for >24h are automatically marked as abandoned on next start
- **Resource registry** -- Indexes installed skills and agents with FTS5 search, composite scoring, and invocation tracking; searchable via `mem_registry` MCP tool
- **Unified resource discovery** -- Shared filesystem traversal layer (`resource-discovery.mjs`) used by both runtime scanner and offline indexer, supporting flat directories, plugin nesting, and loose `.md` files
- **Domain synonym expansion** -- Registry search queries expand to domain synonyms (e.g., "fix" → debug, bugfix, troubleshoot, diagnose, repair)
- **Multi-provider LLM mode** -- Provider priority `ANTHROPIC_API_KEY` (direct Anthropic API) → `OPENROUTER_API_KEY` (OpenRouter, OpenAI-compatible — point it at any model via `OPENROUTER_MODEL`) → `claude -p` CLI fallback when no key is set
- **Lesson-learned indexing** -- `lesson_learned` field indexed in FTS5 with weight 8, making past debugging insights directly searchable
- **Cross-source normalization** -- `mem_search` normalizes scores across observations, sessions, and prompts before merging, preventing any source from dominating results
- **Exponential recency decay** -- Type-differentiated half-lives (decisions: 90d, discoveries: 60d, bugfixes: 14d, changes: 7d) consistently applied in all ranking paths
- **Prompt-time memory injection** -- UserPromptSubmit hook automatically searches and injects relevant past observations with recency and importance weighting
- **Smart skill invocation** -- Auto-loaded and searched managed skills/agents include portable `~` paths with `Read()` guidance; native plugin skills recommend `Skill("full:name")`; prevents `Skill()` misuse for managed resources that aren't registered with Claude Code's native handler
- **Dual injection dedup** -- `user-prompt-search.js` and `handleUserPrompt` coordinate via temp file to prevent duplicate memory injection
- **Plugin cache hook self-heal** -- Claude Code runtime reads plugin hooks from `~/.claude/plugins/cache////hooks/hooks.json`, not from the marketplace source. When `install.mjs`-managed `settings.json` hooks coexist with a stale cache `hooks.json` (e.g. from a previous marketplace install or a plugin auto-update), the runtime registers hooks twice → every session start / user prompt fires twice. `install.mjs` and `hook-update.mjs` now clear cache `hooks.json` in every version dir, and `hook.mjs session-start` self-heals on every session (gated by `hasInstallManagedHooks` so plugin-only users are not affected). `install.mjs status` reports cache pollution state (since v2.31.1/2.31.2).
- **Result-dedup cooldown** -- User-prompt memory injection uses result-overlap detection (>80% ID overlap → skip) instead of time-based cooldown, allowing topic switches within seconds while preventing redundant injections
- **OR query fallback** -- When AND-joined FTS5 queries return zero results, automatically relaxes to OR-joined queries for broader recall (applied in both user-prompt-search and hook-memory paths)
- **Configurable LLM model** -- Switch between Haiku (fast/cheap) and Sonnet (deeper analysis) via `CLAUDE_MEM_MODEL` env var
- **DB auto-recovery** -- Detects and cleans corrupted WAL/SHM files on startup; periodic WAL checkpoints prevent unbounded growth
- **Schema auto-migration** -- Idempotent `ALTER TABLE` migrations run on every startup, safely adding new columns and indexes without data loss
- **Exploration bonus** -- New resources in the registry get a fair chance in composite ranking; zombie resources (high recommend, zero adopt) are penalized in scoring
- **LLM concurrency control** -- File-based semaphore limits background workers to 2 concurrent LLM calls, preventing resource contention
- **stdin overflow protection** -- Hook input truncated at 256KB with regex-based action salvage for oversized tool outputs
- **Cross-session handoff** -- Captures session state (request, completed work, next steps, key files) on `/clear` or `/exit`, then injects context when the next session detects continuation intent via explicit keywords or FTS5 term overlap
- **Git-SHA continuation anchor** (v2.31.0) -- Handoff rows include `git_sha_at_handoff`; any handoff matching the current `HEAD` counts as continuation regardless of TTL. Code state is a stronger continuation signal than wall-clock time
- **Startup dashboard** (v2.31.0) -- SessionStart hook aggregates `git status` + `~/.claude/tasks/*.json` + `~/.claude/plans/*.md` + most-recent exit handoff + recent event count into a single structured block injected via `hookSpecificOutput.additionalContext`
- **Activity namespace** (v2.31.0) -- Dedicated `events` table + FTS5 for non-memdir types (`bugfix`, `lesson`, `bug`, `discovery`, `refactor`, `feature`, `observation`, `decision`) that don't compete with `WHAT_NOT_TO_SAVE` semantics on the observations table. CLI: `claude-mem-lite activity save|search|recent|show`. Slash commands: `/lesson`, `/bug`. `hook-llm` routes non-memdir summary types through `persistHaikuSummary` so upgrades from observations→events are atomic
- **In-place observation updates** -- `mem_update` tool modifies existing observations atomically (field update + FTS text rebuild + vector re-computation in one transaction), preserving original IDs and references
- **Bulk export** -- `mem_export` tool exports observations as JSON or JSONL, with project/type/date filtering and 1000-row pagination cap with batch guidance
- **FTS integrity management** -- `mem_fts_check` tool verifies FTS5 index health or rebuilds indexes on demand, useful after database recovery or when search results seem wrong
- **Atomic multi-table writes** -- `saveObservation` wraps observations + observation_files + observation_vectors INSERTs in a single `db.transaction()`, preventing orphaned rows on crash
- **Modular NLP pipeline** -- Synonym maps, stop words, scoring constants, and query building extracted into focused modules (`synonyms.mjs`, `stop-words.mjs`, `scoring-sql.mjs`, `nlp.mjs`) for independent testing and maintenance
- **Porter-aligned PRF** -- Pseudo-relevance feedback terms are now stemmed with the same Porter algorithm used by FTS5, ensuring PRF expansion terms match the search index
## Platform Support
| Platform | Status | Notes |
|----------|--------|-------|
| **Linux** | Supported | Primary development and testing platform |
| **macOS** | Supported | Fully compatible (Intel and Apple Silicon) |
| **Windows** | Not supported | Uses POSIX shell scripts (`post-tool-use.sh`, `setup.sh`) and Unix file locking; WSL2 may work but is untested |
## Requirements
- **Node.js** >= 18
- **Claude Code** CLI installed and configured (`claude` command available)
- **SQLite3** support (provided by `better-sqlite3`, compiled on install)
- **Platform**: Linux or macOS (see [Platform Support](#platform-support))
## Installation
### Method 1: Plugin Marketplace (recommended)
```bash
/plugin marketplace add sdsrss/claude-mem-lite
/plugin install claude-mem-lite
```
Plugin mode manages its own hooks/runtime. On session start it only **checks and reports** new claude-mem-lite versions; it does **not** self-overwrite plugin files in place. Update plugin-mode installs through Claude's plugin workflow.
> **Auto-adopt fires on the first SessionStart per project (v2.82.1+).** The plugin automatically writes the **invited-memory sentinel** (a system-authority pointer that boosts Claude's proactive use of `mem_recall` / `mem_save`) into the project's memdir — **no manual `/adopt` needed**, regardless of install path (npm, npx, `/plugin`, manual). Per-project opt-out: `claude-mem-lite adopt --disable` (`--enable` to re-arm). Global opt-out: `export MEM_NO_AUTO_ADOPT=1`. Manual `/adopt` remains available for re-applying after edits and for the `--all` batch path.
### Method 2: npx (one-liner)
```bash
npx github:sdsrss/claude-mem-lite
```
Source files are automatically copied to `~/.claude-mem-lite/` for persistence.
### Method 3: git clone
```bash
git clone https://github.com/sdsrss/claude-mem-lite.git
cd claude-mem-lite
node install.mjs install
```
Source files stay in the cloned repo. Update via `git pull && node install.mjs install`.
### What happens during installation
1. **Install dependencies** -- `npm install --omit=dev` (compiles native `better-sqlite3`)
2. **Register MCP server** -- `mem-lite` server with 20 tools (9 core exposed via `tools/list` + 11 hidden-but-callable; see the Usage section for the full table). The pre-v2.78 generic server name `mem` is renamed to `mem-lite` for namespace hygiene; the tool names themselves (`mem_search`, `mem_recall`, ...) are unchanged.
3. **Configure hooks** -- `PostToolUse`, `SessionStart`, `Stop`, `UserPromptSubmit` lifecycle hooks
4. **Create data directory** -- `~/.claude-mem-lite/` (hidden) for database, runtime, and managed resource files
5. **Auto-migrate** -- If `~/.claude-mem/` (original claude-mem) or `~/claude-mem-lite/` (pre-v0.5 unhidden) exists, migrates database and runtime files to `~/.claude-mem-lite/`, preserving the original untouched
6. **Initialize database** -- SQLite with WAL mode, FTS5 indexes created on first server start
Restart Claude Code after installation to activate.
### Migration
All installation methods auto-detect and migrate from previous versions:
**From claude-mem (original `~/.claude-mem/`):**
- Copy `claude-mem.db` → `~/.claude-mem-lite/claude-mem-lite.db` (renamed)
- Copy the `runtime/` directory
- **Original `~/.claude-mem/` is preserved** (no deletion, no overwrite)
**From pre-v0.5 unhidden directory (`~/claude-mem-lite/`):**
- Entire directory is moved to `~/.claude-mem-lite/` (hidden)
**In-place rename:**
- Existing `claude-mem.db` in `~/.claude-mem-lite/` is automatically renamed to `claude-mem-lite.db`
Remove old directories manually after confirming:
```bash
rm -rf ~/.claude-mem/ # original claude-mem
rm -rf ~/claude-mem-lite/ # pre-v0.5 unhidden (if not auto-moved)
```
### Directory Structure
```
~/.claude-mem-lite/
claude-mem-lite.db # SQLite database — memory (WAL mode)
resource-registry.db # SQLite database — skill/agent registry
runtime/
session- # Active session state
ep-.json # Episode buffer
ep-flush-*.json # Flushed episodes awaiting processing
reads-.txt # Read file paths (collected on flush)
managed/
skills/ # Standalone skills: {name}/SKILL.md
agents/ # Agent plugins: {group}/agents/{name}.md + skills/*/SKILL.md
repos/ # Shallow-cloned source repos
```
## Usage
### MCP Tools (used automatically by Claude)
As of v2.70.0, the server registers 20 tools in total but only the 9 **core**
tools appear in `tools/list`. The 11 **hidden** tools remain callable at the
protocol layer (`tools/call` by exact name still routes normally); they're
omitted from the list response so Claude Code sessions don't load 11 extra
tool schemas at startup. Hidden tools are the maintenance / admin / browser
surface — reach them through the CLI column in the second table.
**Core (9, exposed to Claude Code)**
| Tool | Description |
|------|-------------|
| `mem_search` | FTS5 full-text search with BM25 ranking. Filters by type, project, date range, importance level. |
| `mem_recent` | Show most recent observations, ordered by time. Quick snapshot of latest activity. |
| `mem_recall` | Recall observations related to a file. Use before editing to surface past bugfixes and context. |
| `mem_timeline` | Browse observations chronologically around an anchor point. |
| `mem_get` | Retrieve full details for specific observation IDs (includes importance and related_ids). |
| `mem_save` | Manually save a memory/observation. Accepts `closes_deferred` array for transactional closure of deferred work. |
| `mem_defer` | Mark work for a future session (v2.70+). First-class carry-forward signal, surfaced in SessionStart `### Deferred Work` block. |
| `mem_defer_list` | List open deferred items for the current project. |
| `mem_defer_drop` | Drop a deferred item without fixing it; requires a `reason` for the audit trail. |
**Hidden-but-callable (11, CLI-routed)**
| Tool | CLI equivalent | Notes |
|------|----------------|-------|
| `mem_update` | `claude-mem-lite update ` | Edit an observation in place. |
| `mem_stats` | `claude-mem-lite stats` | Counts, type distribution, daily activity. |
| `mem_delete` | `claude-mem-lite delete ` | Preview / confirm workflow, FTS5 cleanup. |
| `mem_compress` | `claude-mem-lite compress --preview` | Roll up old low-value observations. |
| `mem_maintain` | `claude-mem-lite maintain --action scan` | dedup / decay / cleanup / rebuild_vectors. |
| `mem_optimize` | `claude-mem-lite optimize --action preview` | LLM-powered re-enrich / normalize / cluster-merge. |
| `mem_export` | `claude-mem-lite export` | JSON / JSONL dump, filters by project, type, date. |
| `mem_fts_check` | `claude-mem-lite fts-check [--rebuild]` | FTS5 integrity + rebuild. |
| `mem_browse` | `claude-mem-lite browse` | Tier-grouped dashboard (working / active / archive). |
| `mem_registry` | `claude-mem-lite registry ` | List / search / import / remove skills + agents. |
| `mem_use` | _MCP only_ | Load a skill / agent from the registry by name. |
### Skill Commands (in Claude Code chat)
```
/mem search # Full-text search across all memories
/mem recent [n] # Show recent N observations (default 10)
/mem recall # Show past observations for a file
/mem save # Save a manual memory/note
/mem stats # Show memory statistics
/mem timeline # Browse timeline around a match
/mem browse # Tier-grouped memory dashboard
/mem # Shorthand for search
/lesson # Save a non-obvious lesson to the events table (v2.31.0)
/bug # Log a known bug + repro steps to the events table (v2.31.0)
```
### Efficient Search Workflow
```
1. mem_search(query="auth bug") -> compact ID index
2. mem_timeline(anchor=12345) -> surrounding context
3. mem_get(ids=[12345, 12346]) -> full details
```
### Invited Memory (v2.32+)
Opt-in mechanism that installs a single sentinel-wrapped line into the project's
memdir (`~/.claude/projects//memory/MEMORY.md`) so Claude Code loads
the plugin's MCP-tool triggers as **user-memory** — a higher instruction-following
authority than MCP server instructions (which are framed as tool metadata).
```bash
claude-mem-lite adopt # install for current project
claude-mem-lite adopt --all # install for every project under ~/.claude/projects/
claude-mem-lite adopt --status # list adopted/disabled projects + current gating snapshot
claude-mem-lite adopt --dry-run # preview without writing
claude-mem-lite adopt --disable # opt out of auto-adopt for current project (writes .mem-no-auto-adopt sentinel)
claude-mem-lite adopt --enable # re-arm auto-adopt for current project (deletes the sentinel)
claude-mem-lite unadopt # remove sentinel + doc (runtime marker stays to honor the explicit removal)
```
Slash commands `/adopt` and `/unadopt` wrap the same CLI.
**What adoption changes:**
- A `…` block is
added to `MEMORY.md` under a `## 插件契约` header, containing one ≤150-char
line pointing at `mem_recall` / `mem_save` with their key arguments.
- A `plugin_claude_mem_lite.md` detail file is written (not auto-loaded; read
on demand when the MEMORY.md pointer surfaces in context).
- Post-adopt the conservative hook layer auto-trims: MCP server instructions
drop the `WHEN TO USE` section, SessionStart injection drops the `File Lessons`
/ `Key Context` sections. `#ID` references and the `Recent` table still fire
so `mem_get` remains reachable.
**When does it take effect?**
- The MEMORY.md sentinel and the hook-layer trim (`File Lessons` / `Key Context` /
lesson suffix) apply on the **next SessionStart** (any new Claude Code session
in the adopted project).
- The MCP server instructions are built once at server boot and MCP has no
"push" protocol — the `WHEN TO USE` / `Decision rules` trim only applies
after Claude Code restarts and re-spawns the mem-lite MCP server. A single
`/exit` + fresh session is enough. Same caveat applies to `unadopt`.
**Safety:**
- Hash-guarded: editing the sentinel body yourself blocks automatic rewrites
unless you pass `--force`.
- Budget-gated: refuses to insert when MEMORY.md is already >180 lines, so
Claude Code's 200-line MEMORY.md cap won't truncate the block.
- **Auto-adopt fires on the first SessionStart per project for any install
path (v2.82.1+).** Per-project opt-out: `claude-mem-lite adopt --disable`
(writes a durable `/.mem-no-auto-adopt` sentinel that survives marker
deletion / plugin reinstalls). Global opt-out: `MEM_NO_AUTO_ADOPT=1`.
Pre-v2.82.1 the `CLAUDE_PLUGIN_ROOT` gate left auto-adopt unreachable for
every `install.mjs`-written hook (the common path) — see CHANGELOG v2.82.1.
- The fallback hook layer is never removed from source — conditional trim is
runtime-gated on sentinel presence, so projects without adoption get the
full verbose output.
See `docs/plans/2026-04-16-invited-memory-pattern.md` for the full design
(including the reusable template other plugins can follow).
## Database Schema
Five core tables with FTS5 virtual tables for search:
**observations** -- Individual coding observations (decisions, bugfixes, features, etc.)
```
id, memory_session_id, project, type, title, subtitle,
text, narrative, concepts, facts, files_read, files_modified,
importance, related_ids, created_at, created_at_epoch,
lesson_learned, minhash_sig, access_count, compressed_into, search_aliases,
branch, superseded_at, superseded_by, last_accessed_at
```
**session_summaries** -- LLM-generated session summaries
```
id, memory_session_id, project, request, investigated,
learned, completed, next_steps, files_read, files_edited, notes,
remaining_items, lessons, key_decisions
```
**sdk_sessions** -- Session tracking
```
id, content_session_id, memory_session_id, project,
started_at, completed_at, status, prompt_counter
```
**user_prompts** -- User prompts captured via UserPromptSubmit hook
```
id, content_session_id, prompt_text, prompt_number
```
**session_handoffs** -- Cross-session handoff snapshots (UPSERT, max 2 per project)
```
project, type, session_id, working_on, completed, unfinished,
key_files, key_decisions, match_keywords, created_at_epoch
```
**observation_files** -- Normalized file membership for efficient file-based recall
```
obs_id, filename
```
**observation_vectors** -- TF-IDF vector embeddings for hybrid search
```
observation_id, vector (BLOB Float32Array), vocab_version, created_at_epoch
```
**vocab_state** -- Persisted TF-IDF vocabulary for stable vector indexing
```
term, term_index, idf, version, created_at_epoch
```
FTS5 indexes: `observations_fts` (title, subtitle, narrative, text, facts, concepts, lesson_learned), `session_summaries_fts`, `user_prompts_fts`
## How It Works
### Hook Pipeline
```
SessionStart
-> Generate session ID (or save handoff snapshot on /clear)
-> Mark stale sessions (>24h active) as abandoned
-> Clean orphaned/stale lock files
-> Query recent observations (24h)
-> Inject context into CLAUDE.md + stdout
PostToolUse (every tool execution)
-> Bash pre-filter skips noise in ~5ms (Read paths tracked to reads file)
-> Detect Bash significance (errors, tests, builds, git, deploys)
-> Accumulate into episode buffer
-> Proactive file history: show past observations for edited files
-> Flush when: buffer full (10 entries) | 5min gap | context change
-> Collect Read file paths into episode on flush
-> Spawn LLM episode worker for significant episodes
-> Error-triggered recall: search memory for related past fixes
UserPromptSubmit (two parallel paths)
-> [user-prompt-search.js] Auto-search memory via FTS5 + active file context
-> [user-prompt-search.js] Inject relevant past observations with recency/importance weighting
-> [user-prompt-search.js] Write injected IDs to temp file for dedup
-> [user-prompt-search.js] L1 skill auto-load: match managed skill names in prompt
-> Load content with portable ~ path + Read() guidance
-> source="managed-skill|managed-agent", path="~/.claude-mem-lite/managed/..."
-> [hook.mjs handleUserPrompt] Capture user prompt text to user_prompts table
-> [hook.mjs handleUserPrompt] Increment session prompt counter
-> [hook.mjs handleUserPrompt] Handoff: detect continuation intent → inject previous session context
-> [hook.mjs handleUserPrompt] Semantic memory injection (hook-memory.mjs), deduped via temp file
Stop
-> Flush final episode buffer
-> Save handoff snapshot (on /exit)
-> Mark session completed
-> Spawn LLM summary worker (poll-based wait)
```
### Resource Registry
The resource registry (`registry.mjs`, `registry-retriever.mjs`) indexes installed skills and agents into a searchable FTS5 database. Unlike the previous proactive dispatch system, the registry is now on-demand — it's reachable via the `claude-mem-lite registry` CLI (primary path for Claude Code since v2.34.0 hides the `mem_registry` MCP tool from `tools/list`) or by direct `tools/call mem_registry` for MCP clients that know the name.
```
Registry pipeline:
-> registry-scanner.mjs discovers skills/agents on filesystem
-> resource-discovery.mjs handles flat dirs, plugin nesting, loose .md files
-> registry-indexer.mjs indexes content into FTS5 with metadata
-> registry-retriever.mjs provides BM25-ranked search with synonym expansion
-> mem_registry MCP tool exposes search/list/stats/import/remove/reindex actions
Smart invocation (three layers):
L1 auto-load: UserPromptSubmit matches managed skill name in prompt
-> Loads content with path="~/.claude-mem-lite/managed/.../SKILL.md"
-> Guides: Read("path") or mem_use(name="..."), never Skill()
L2 bridge: PreToolUse hook intercepts Skill("name") for managed resources
-> Outputs content, prevents native handler failure
L3 explicit: mem_use(name="...") loads full content with reload path
Search: managed resources → Read(path), native plugins → Skill("full:name")
```
Composite scoring for search results: BM25 relevance (40%) + repo stars (15%) + success rate (15%) + adoption rate (10%) + freshness (10%) + exploration bonus (10%). Domain filtering ensures platform-specific resources (iOS, Go, Rust) only surface for matching projects.
### Episode Encoding
Episodes are batched related operations (edits to the same file group) that get processed by a background LLM worker:
```
Episode buffer -> Flush to JSON -> claude -p --model haiku -> Structured observation -> SQLite
```
Each observation includes type, title, narrative, concepts, facts, importance (1-3), and is automatically deduplicated via two tiers: Jaccard similarity (>70% within 5 minutes) and MinHash signatures (>80% within 7 days across sessions). If the LLM call fails, a degraded observation is saved with inferred metadata (zero data loss). Related observations are linked via `related_ids` based on FTS5 title similarity and file overlap.
## Management Commands
```bash
# Plugin install:
/plugin install claude-mem-lite # Install / update
/plugin uninstall claude-mem-lite # Uninstall
# git clone install:
node install.mjs install # Install and configure
node install.mjs uninstall # Remove (keep data)
node install.mjs uninstall --purge # Remove and delete all data
node install.mjs status # Show current status
node install.mjs doctor # Diagnose issues
node install.mjs cleanup-hooks # Remove only stale claude-mem-lite hooks from settings.json
node install.mjs update # Force-check for updates and install them (direct install / npx mode)
# npx install:
npx claude-mem-lite # Install / reinstall
npx claude-mem-lite uninstall # Remove (keep data)
npx claude-mem-lite doctor # Diagnose issues
```
Notes:
- Plugin mode only reports available updates; it does not self-update plugin files.
To upgrade an installed plugin to the latest published version, run **inside Claude Code**:
```
/plugin marketplace update sdsrss
/plugin install claude-mem-lite@sdsrss
```
(The first command refreshes the local marketplace clone; the second reinstalls from it. Without the first command, `/plugin install` reuses the stale local clone and you stay on whichever version you originally pulled.)
- Direct install / npx mode keeps auto-update enabled and uses staged replacement with rollback on install failure.
- If you disabled the plugin but still have old mem hooks in `~/.claude/settings.json`, run `node install.mjs cleanup-hooks`.
### doctor
Checks Node.js version, dependencies, server/hook files, database integrity, FTS5 indexes, and stale processes.
### status
Shows MCP registration, hook configuration, plugin disabled state, and database stats (observation/session counts).
### Recovery (stuck install / hook errors)
If you see `ERR_MODULE_NOT_FOUND` on PreToolUse:Read/Edit/Skill hooks, or `claude-mem-lite` commands crash with import errors, you're likely hit by a partial auto-update — the updater copied new scripts but missed a sibling `lib/*` file, breaking the hook chain (and the next auto-update that would have healed it).
**v2.84.0+** ships a `repair` subcommand that re-syncs from the latest GitHub release:
```bash
claude-mem-lite repair
```
**If `repair` itself fails** (the bin is older than v2.84.0, or the bin is also broken), run this one-liner — it pulls a fresh tarball into a temp dir and runs *that* tarball's `install.mjs`, bypassing every file on your disk:
```bash
T=$(mktemp -d) && curl -sL https://api.github.com/repos/sdsrss/claude-mem-lite/tarball | tar xz -C "$T" --strip-components=1 && node "$T/install.mjs" install
```
After it finishes, `~/.claude-mem-lite/` is back in sync with the latest release and `claude-mem-lite repair` is available for next time.
## Uninstall
```bash
# Plugin:
/plugin uninstall claude-mem-lite
# git clone:
cd claude-mem-lite
node install.mjs uninstall # Keeps ~/.claude-mem-lite/ data
node install.mjs uninstall --purge # Deletes ~/.claude-mem-lite/ and all data
# npx:
npx claude-mem-lite uninstall
npx claude-mem-lite uninstall --purge
```
Data in `~/.claude-mem-lite/` is preserved by default. Delete manually if needed:
```bash
rm -rf ~/.claude-mem-lite/
```
### Mixed-install residue (read this if you've used multiple install methods)
`/plugin uninstall` only removes the plugin manifest — it **does not touch `~/.claude/settings.json`**. If you've ever run `claude-mem-lite install` (npx or git-clone path), hook entries pointing at `~/.claude-mem-lite/hook.mjs` were written into your user-global settings, and they keep firing after `/plugin uninstall`. If `~/.claude-mem-lite/hook.mjs` still exists they double-fire alongside the plugin; if you also ran `rm -rf ~/.claude-mem-lite/` they error every session.
**The safe sequence is**: run `claude-mem-lite uninstall` first (which cleans the settings.json hooks plus the global MCP registration), then `/plugin uninstall claude-mem-lite`, then optionally `rm -rf ~/.claude-mem-lite/`.
If you already uninstalled in the wrong order, `claude-mem-lite doctor` flags orphan hooks under `Orphan hooks:` with the exact cleanup command.
## Project Structure
```
claude-mem-lite/
.claude-plugin/
plugin.json # Plugin manifest
marketplace.json # Marketplace catalog
.mcp.json # MCP server definition (plugin root)
hooks/
hooks.json # Hook definitions (plugin mode)
commands/
mem.md # /mem command definition
server.mjs # MCP server: tool definitions, FTS5 search, database init
server-internals.mjs # Extracted search helpers: re-ranking, PRF, concept expansion
hook.mjs # Claude Code hooks: episode capture, error recall, session management
hook-llm.mjs # Background LLM workers: episode extraction, session summaries
hook-shared.mjs # Shared hook infrastructure: session management, DB access, LLM calls
hook-handoff.mjs # Cross-session handoff: state extraction, intent detection, injection
hook-context.mjs # CLAUDE.md context injection and token budgeting
hook-episode.mjs # Episode buffer management: atomic writes, pending entry merging
hook-memory.mjs # Semantic memory injection on user prompt
hook-semaphore.mjs # LLM concurrency control: file-based semaphore for background workers
schema.mjs # Database schema: single source of truth for tables, migrations, FTS5
tool-schemas.mjs # Shared Zod schemas for MCP tool validation
tfidf.mjs # TF-IDF vector engine: tokenization, vocabulary building, vector computation, cosine similarity, RRF merge
tier.mjs # Temporal tier system: activity-based time window classification
utils.mjs # Re-export hub: backward-compatible surface for all utility modules
nlp.mjs # FTS5 query building: synonym expansion, CJK bigrams, sanitization
scoring-sql.mjs # BM25 weight constants and type-differentiated decay half-lives
stop-words.mjs # Shared base stop-word set for all NLP/search modules
synonyms.mjs # Unified synonym source: SYNONYM_MAP (bidirectional) + DISPATCH_SYNONYMS
project-utils.mjs # Shared project name resolution with in-process cache
secret-scrub.mjs # API key, token, PEM, and credential pattern redaction
format-utils.mjs # String formatting: truncate, typeIcon, date/time/week formatting
hash-utils.mjs # MinHash signatures, Jaccard similarity for dedup
bash-utils.mjs # Bash output significance detection: errors, tests, builds, deploys
# Resource registry
registry.mjs # Resource registry DB: schema, CRUD, FTS5, invocation tracking
registry-retriever.mjs # FTS5 retrieval with synonym expansion and composite scoring
registry-indexer.mjs # Resource indexing pipeline
registry-scanner.mjs # Filesystem scanner: reads content + hashes, delegates discovery
resource-discovery.mjs # Shared discovery layer: flat dirs, plugin nesting, loose .md files
haiku-client.mjs # Unified Haiku LLM wrapper: direct API or CLI fallback
# Install & config
install.mjs # CLI installer: setup, uninstall, status, doctor (npx/git clone mode)
skill.md # MCP skill definition (npx/git clone mode)
package.json # Dependencies and metadata
scripts/
setup.sh # Setup hook: npm install + migration (hidden dir + old dir)
post-tool-use.sh # Bash pre-filter: skips noise in ~5ms, tracks Read paths
user-prompt-search.js # UserPromptSubmit hook: auto-search memory + L1 skill auto-load
pre-skill-bridge.js # PreToolUse hook: L2 skill bridge for managed resources
pre-tool-recall.js # PreToolUse hook: file lesson recall before Edit/Write
prompt-search-utils.mjs # Shared logic: skip patterns, intent detection, name matching
convert-commands.mjs # Converts command .md → SKILL.md in managed plugins
index-managed.mjs # Offline indexer for managed resources
# Test & benchmark (dev only)
tests/ # Unit, property, integration, contract, E2E, pipeline tests
benchmark/ # BM25 search quality benchmarks + CI gate
```
## Search Quality
Benchmarked on 200 observations across 30 queries (standard + hard-negative categories):
| Metric | Score |
|--------|-------|
| Recall@10 | 0.88 |
| Precision@10 | 0.96 |
| nDCG@10 | 0.95 |
| MRR@10 | 0.95 |
| P95 search latency | 0.15ms |
The benchmark suite runs as a CI gate (`npm run benchmark:gate`) to prevent search quality regressions.
## Development
```bash
npm run lint # ESLint static analysis
npm test # Run full test suite (vitest)
npm run test:smoke # Run 5 core smoke tests
npm run test:coverage # Run tests with V8 coverage (≥75% lines/functions, ≥65% branches)
npm run benchmark # Run full search quality benchmark
npm run benchmark:gate # CI gate: fails if metrics regress beyond 5% tolerance
```
## Environment Variables
| Variable | Description | Default |
|----------|-------------|---------|
| `CLAUDE_MEM_DIR` | Custom data directory. All databases, runtime files, and managed resources are stored here. | `~/.claude-mem-lite/` |
| `CLAUDE_MEM_MODEL` | LLM model for background calls (episode extraction, session summaries). Accepts `haiku` or `sonnet`. | `haiku` |
| `ANTHROPIC_API_KEY` | Anthropic API key. When set, all background LLM calls go directly to the Anthropic Messages API (with prompt caching). Highest priority. | _(unset → CLI)_ |
| `OPENROUTER_API_KEY` | OpenRouter API key (OpenAI-compatible). Used for background LLM calls when `ANTHROPIC_API_KEY` is **not** set. If neither key is set, calls fall back to the `claude -p` CLI. | _(unset)_ |
| `OPENROUTER_MODEL` | Overrides the OpenRouter model slug for **all** background calls (e.g. `openai/gpt-4o-mini`, `qwen/qwen-2.5-72b-instruct`). When unset, the `CLAUDE_MEM_MODEL` tier maps to `anthropic/claude-haiku-4.5` (haiku) or `anthropic/claude-sonnet-4.5` (sonnet). | _(tier default)_ |
| `CLAUDE_MEM_DEBUG` | Enable debug logging (`1` to enable). | _(disabled)_ |
| `MEM_QUIET_HOOKS` | Low-noise hooks. `1` drops the `File Lessons` / `Key Context` sections from SessionStart injection, the lesson suffix from `[mem] Related memories`, and the `WHEN TO USE` / `Decision rules` blocks from MCP server instructions. IDs and the `Recent` table still surface so `mem_get(ids=[…])` remains reachable. Intended for users running the invited-memory adopt path or who otherwise want minimal auto-injection. **Since v2.82.0 this env no longer gates auto-adopt — use `MEM_NO_AUTO_ADOPT=1` for that.** | _(disabled)_ |
| `MEM_NO_AUTO_ADOPT` | Global opt-out for auto-adopt (v2.82.0+). `1` prevents the first-SessionStart auto-write of the invited-memory sentinel across **all** projects. For per-project opt-out use `claude-mem-lite adopt --disable` instead (writes a durable `/.mem-no-auto-adopt` sentinel that survives marker deletion). | _(disabled)_ |
| `MEM_NO_ADOPT_HINT` | Silences the one-line "Invited-memory 未启用：`claude-mem-lite adopt`…" hint that SessionStart appends when the current project hasn't been adopted. Since v2.82.1 auto-adopt fires on first SessionStart for any install path, so this hint typically surfaces only when you've explicitly opted out (`MEM_NO_AUTO_ADOPT=1` or `claude-mem-lite adopt --disable`). | _(disabled)_ |
## FAQ
### What is a memory system for Claude Code?
A memory system lets Claude Code remember context — coding decisions, bug fixes, file history — across sessions. By default Claude Code's context resets each session; claude-mem-lite persists observations to a local SQLite database and re-injects them at session start and on relevant prompts.
### Does Claude Code have built-in long-term memory?
No. Claude Code's `CLAUDE.md` and `MEMORY.md` files act as static instruction memory, but there is no native dynamic recall of past sessions, bug fixes, or decisions. claude-mem-lite adds that layer via MCP and hooks, with no manual note-taking required.
### How is claude-mem-lite different from mem0 or MCP's reference memory server?
`mem0` and the MCP `memory` server are general-purpose LLM memory frameworks designed for any client. claude-mem-lite is purpose-built for Claude Code's hook lifecycle: it captures *episodes* (batched tool calls), uses domain-specific synonym expansion for code terms (`K8s`, `DB`, `数据库`, ...), and surfaces past observations proactively before file edits via the `PreToolUse:Edit` hook.
### Why "lite"? What did the original claude-mem do differently?
The original called an LLM on every tool use with raw JSON inputs. claude-mem-lite batches 5–10 operations per LLM call, uses a smaller model (Haiku), and runs a deterministic code-level filter before sending anything to the model. Net result: an estimated ~600× lower cost (an architecture estimate from the cost model above, not a measured benchmark) with equivalent search quality. See the [Architecture comparison](#architecture-comparison) above.
### Does this work cross-project? Cross-machine?
Project-scoped by default — each project has its own memory namespace. Single-machine only (SQLite, not networked). Use `mem_export` (JSON / JSONL) to back up or migrate between machines.
### What about privacy? Does it call external APIs?
Only the Haiku summarization step calls Anthropic's API (or the local `claude -p` CLI if no API key is set). All search, storage, and retrieval is local SQLite — no telemetry, no third-party services.
### 中文常见问题
**Claude Code 怎么跨会话记住内容？** 默认不能。claude-mem-lite 通过 MCP 协议和钩子自动把决策、bug 修复、文件历史持久化到本地 SQLite，下次会话开始时再注入。
**和 mem0、官方 MCP memory server 有什么区别？** 那两个是通用 LLM 记忆框架；claude-mem-lite 是为 Claude Code 钩子生命周期定制的：批量 episode 处理、代码领域同义词扩展（K8s/DB/数据库等）、文件编辑前主动召回相关历史。
**支持中文吗？** 完整支持。FTS5 + 中英文同义词扩展（100+ 对，含 CJK ↔ EN 跨语言映射），中文记忆也可用英文关键词召回，反之亦然。
## License
MIT