ClawMem — On-device memory layer for Claude Code, OpenClaw, and Hermes agents

On-device memory for Claude Code, OpenClaw, Hermes, and AI agents. Retrieval-augmented search, hooks, and an MCP server in a single local system. No API keys, no cloud dependencies.

ClawMem fuses recent research into a retrieval-augmented memory layer that agents actually use. The hybrid architecture combines QMD-derived multi-signal retrieval (BM25 + vector search + reciprocal rank fusion + query expansion + cross-encoder reranking), SAME-inspired composite scoring (recency decay, confidence, content-type half-lives, co-activation reinforcement), MAGMA-style intent classification with multi-graph traversal (semantic, temporal, and causal beam search), and A-MEM self-evolving memory notes that enrich documents with keywords, tags, and causal links between entries. Pattern extraction from Engram adds deduplication windows, frequency-based durability scoring, and temporal navigation.

Integrates via Claude Code hooks, an MCP server (works with any MCP-compatible client), a native OpenClaw plugin, or a Hermes Agent MemoryProvider plugin. All paths write to the same local SQLite vault. A decision captured during a Claude Code session shows up immediately when an OpenClaw or Hermes agent picks up the same project.

TypeScript on Bun. MIT License.

What It Does

ClawMem turns your markdown notes, project docs, and research dumps into persistent memory for AI coding agents. It automatically:

• Surfaces relevant context on every prompt (context-surfacing hook)
• Bootstraps sessions with your profile, latest handoff, recent decisions, and stale notes
• Captures decisions, preferences, milestones, and problems from session transcripts using a local GGUF observer model
• Imports conversation exports from Claude Code, ChatGPT, Claude.ai, Slack, and plain text via clawmem mine, with optional post-import LLM fact extraction (--synthesize) that pulls structured decisions / preferences / milestones / problems and cross-fact links out of otherwise full-text conversation dumps (v0.7.2)
• Generates handoffs at session end so the next session can pick up where you left off
• Learns what matters via a feedback loop that boosts referenced notes and decays unused ones
• Guards against prompt injection in surfaced content
• Classifies query intent (WHY / WHEN / ENTITY / WHAT) to weight search strategies
• Traverses multi-graphs (semantic, temporal, causal) via adaptive beam search
• Evolves memory metadata as new documents create or refine connections
• Infers causal relationships between facts extracted from session observations
• Detects contradictions between new and prior decisions, auto-decaying superseded ones (with an additional merge-time contradiction gate in the consolidation worker that blocks cross-observation contradictions before they land, v0.7.1)
• Guards against cross-entity merges during consolidation — name-aware dual-threshold merge safety compares entity anchors before merging similar observations, preventing "Alice decided X" from merging into "Bob decided X" (v0.7.1)
• Prevents context bleed in derived insights — the Phase 3 deductive synthesis pipeline validates every draft against an anti-contamination wrapper (deterministic entity contamination check + LLM validator + dedupe) before writing cross-session deductive observations (v0.7.1)
• Frames surfaced facts as background knowledge — context-surfacing wraps injected content in <instruction> + <facts> + <relationships> blocks, telling the model to treat facts as already-known and exposing memory-graph edges between surfaced docs directly in-prompt (v0.7.1)
• Injects knowledge-graph facts as structured triples — when the user's prompt mentions entities already known to the vault, context-surfacing resolves them via a three-path prompt-only extractor (canonical IDs, proper nouns, lowercased n-grams), queries the SPO graph for current-state triples, and appends a <vault-facts> block of raw subject predicate object lines to <vault-context> — off for speed, 200 tokens on balanced, 250 on deep, token-truncated at the triple boundary (v0.9.0)
• Session-scoped focus topic boost — clawmem focus set "<topic>" --session-id <id> writes a per-session focus file that steers query expansion, reranking, chunk selection, snippet extraction, and post-composite-score topic boosting (1.4× match / 0.75× demote) for that session only — session-isolated, fail-open, never writes to SQLite or lifecycle columns (v0.9.0)
• Scores document quality using structure, keywords, and metadata richness signals
• Boosts co-accessed documents — notes frequently surfaced together get retrieval reinforcement
• Decomposes complex queries into typed retrieval clauses (BM25/vector/graph) for multi-topic questions
• Cleans stale embeddings automatically before embed runs, removing orphans from deleted/changed documents
• Transaction-safe indexing — crash mid-index leaves zero partial state (atomic commit with rollback)
• Deduplicates hook-generated observations within a 30-minute window using normalized content hashing, preventing memory bloat from repeated hook output
• Navigates temporal neighborhoods around any document via the timeline tool — progressive disclosure from search to chronological context to full content
• Boosts frequently-revised memories — documents with higher revision counts get a durability signal in composite scoring (capped at 10%)
• Supports pin/snooze lifecycle for persistent boosts and temporary suppression
• Manages document lifecycle — policy-driven archival sweeps with restore capability
• Auto-routes queries via memory_retrieve — classifies intent and dispatches to the optimal search backend
• Syncs project issues from Beads issue trackers into searchable memory
• Runs a quiet-window heavy maintenance lane — a second consolidation worker, off by default behind CLAWMEM_HEAVY_LANE=true, that runs on a longer interval only inside a configurable hour window. Gated by context_usage query-rate so it never competes for CPU/GPU with interactive sessions, scoped exclusively via DB-backed worker_leases, stale-first by default with an optional surprisal selector, and journals every attempt in maintenance_runs for operator visibility (v0.8.0)

Runs fully local with no API keys and no cloud services. Integrates via Claude Code hooks and MCP tools, as a native OpenClaw plugin, or as a Hermes Agent MemoryProvider plugin. All modes share the same vault for cross-runtime memory. Works with any MCP-compatible client.

Full version history is in RELEASE_NOTES.md. Upgrade instructions for existing vaults are in docs/guides/upgrading.md.

Architecture

Install

Platform Support

Prerequisites

Required:

• Bun v1.0+ — runtime for ClawMem. On Linux, install via curl -fsSL https://bun.sh/install | bash (not snap — snap Bun cannot read stdin, which breaks hooks).
• SQLite with FTS5 — included with Bun. On macOS, install brew install sqlite for extension loading support (ClawMem detects and uses Homebrew SQLite automatically).

Optional (for better performance):

• llama.cpp (llama-server) — for dedicated GPU inference. Without it, node-llama-cpp runs models in-process (auto-downloads on first use). GPU servers give better throughput and prevent silent CPU fallback.
• systemd (Linux) or launchd (macOS) — for persistent background services (watcher, embed timer, GPU servers). ClawMem ships systemd unit templates; macOS users can create equivalent launchd plists. See systemd services.

Optional integrations:

• Claude Code — for hooks + MCP integration
• OpenClaw — for native plugin integration
• Hermes Agent — for MemoryProvider plugin integration
• bd CLI v0.58.0+ — for Beads issue tracker sync (only if using Beads)

Install from npm (recommended)

npm install -g clawmem

If you use Bun as your package manager:

bun add -g clawmem

Install from source

git clone https://github.com/yoloshii/clawmem.git ~/clawmem
cd ~/clawmem && bun install
ln -sf ~/clawmem/bin/clawmem ~/.bun/bin/clawmem

Setup roadmap

After installing, here's the full journey from zero to working memory:

Fastest path: Step 1 alone gets you a working system with in-process CPU/GPU inference and default models — no manual model downloads or service configuration needed. Steps 2-4 are optional upgrades for better performance. Steps 5-6 are where you customize what gets indexed and how your agent connects.

Customize what gets indexed: Each collection has a pattern field in ~/.config/clawmem/config.yaml (default: **/*.md). Tailor it per collection — index project docs, research notes, decision records, Obsidian vaults, or anything else your agents should know about. The more relevant content in the vault, the better retrieval works. See the quickstart for config examples.

Quick start commands

# One command: init + index + embed + hooks + MCP
clawmem bootstrap ~/notes --name notes

# Or step by step:
clawmem init
clawmem collection add ~/notes --name notes
clawmem update --embed
clawmem setup hooks
clawmem setup mcp

# Add more collections (the more you index, the richer retrieval gets)
clawmem collection add ~/projects/myapp --name myapp
clawmem collection add ~/research --name research
clawmem update --embed

# Verify
clawmem doctor

Upgrading

bun update -g clawmem   # or: npm update -g clawmem

Database schema migrates automatically on next startup (new tables and columns are added via CREATE IF NOT EXISTS / ALTER TABLE ADD COLUMN).

After major version updates (e.g. 0.1.x → 0.2.0) that add new enrichment pipelines, run a full enrichment pass to backfill existing documents:

clawmem reindex --enrich  # Full enrichment: entity extraction + links + evolution for all docs
clawmem embed             # Re-embed if upgrading embedding models (not needed for most updates)

--enrich forces the complete A-MEM pipeline (entity extraction, link generation, memory evolution) on all documents, not just new ones. Without it, reindex only refreshes metadata for existing docs.

Routine patch updates (e.g. 0.2.0 → 0.2.1) do not require reindexing.

For version-specific upgrade notes (opt-in features, optional cleanup steps, verification commands), see docs/guides/upgrading.md.

Integration

Claude Code

ClawMem integrates via hooks (settings.json) and an MCP stdio server. Hooks handle 90% of retrieval automatically - the agent never needs to call tools for routine context.

clawmem setup hooks    # Install lifecycle hooks (SessionStart, UserPromptSubmit, Stop, PreCompact)
clawmem setup mcp      # Register MCP server in ~/.claude.json (31 tools)

Automatic (90%): context-surfacing injects relevant memory on every prompt. postcompact-inject re-injects state after compaction. decision-extractor, handoff-generator, feedback-loop capture session state on stop.

Agent-initiated (10%): MCP tools (query, intent_search, find_causal_links, timeline, etc.) for targeted retrieval when hooks don't surface what's needed.

OpenClaw

ClawMem registers as a native OpenClaw memory plugin (kind: memory, v0.10.0+). Same 90/10 automatic retrieval, delivered through OpenClaw's plugin-hook bus instead of Claude Code hooks.

clawmem setup openclaw   # Installs plugin into ~/.openclaw/extensions/clawmem (copy, not symlink)

What the plugin provides:

• before_prompt_build hook (load-bearing) - prompt-aware retrieval (context-surfacing + session-bootstrap) AND the pre-emptive precompact-extract run when token usage approaches the compaction threshold. This is the authoritative path for precompact state capture because it runs synchronously before the LLM call that would trigger compaction, so it cannot race the compactor.
• agent_end hook - decision extraction, handoff generation, feedback loop (parallel, fire-and-forget at the OpenClaw call site)
• before_compaction hook (defense-in-depth fallback) - fires precompact-extract again for the rare case where before_prompt_build's proximity heuristic missed a sudden token-count jump. Fire-and-forget at OpenClaw's call site, so it races the compactor and offers no correctness guarantee on its own — the before_prompt_build path is what actually holds the invariant.
• session_start hook - session registration + cached first-turn bootstrap context
• 5 agent tools - clawmem_search, clawmem_get, clawmem_session_log, clawmem_timeline, clawmem_similar

Disable OpenClaw's native memory search to avoid duplicate injection:

openclaw config set agents.defaults.memorySearch.extraPaths "[]"

ClawMem coexists cleanly with OpenClaw's Active Memory plugin (v2026.4.10+). They search different backends and inject into different prompt regions, so both can run simultaneously without conflict. See the OpenClaw plugin guide for details.

Pair ClawMem (memory) with a context-engine plugin (v0.10.0+). OpenClaw and Hermes maintainers have converged on a two-surface plugin model: one slot for memory plugins (cross-session, retrieval-first) and a separate slot for context-engine plugins (in-session, compression/compaction-first). Under that model ClawMem is a memory layer — it has always been one in Hermes via the MemoryProvider ABC, and v0.10.0 moves the OpenClaw integration to the same semantic slot. You can now run ClawMem in the memory slot alongside an LCM-style compression plugin (for example, lossless-claw) in the context-engine slot. The two plugins do not overlap: one persists across sessions, the other reshapes the live window. See the OpenClaw plugin guide — memory vs context engine for the full rationale.

OpenClaw v2026.4.11+ recommended (required for ClawMem v0.10.0+). v2026.4.11 introduced a new plugin discovery contract that requires each plugin directory to ship a package.json with openclaw.extensions declared, and that rejects symlinked plugin directories. ClawMem v0.10.0 includes both fixes. Older ClawMem versions (< v0.10.0) on OpenClaw v2026.4.11+ will fail to discover silently — upgrade ClawMem, then re-run clawmem setup openclaw. See docs/guides/upgrading.md.

Alternative: OpenClaw agents can also use ClawMem's MCP server directly (clawmem setup mcp), with or without hooks. This gives full access to all 31 MCP tools but bypasses OpenClaw's plugin lifecycle, so you lose token budget awareness, native compaction orchestration, and the agent_end message pipeline. The native OpenClaw plugin is recommended for new setups; MCP is available as an additional or standalone integration.

Hermes Agent

ClawMem integrates as a native MemoryProvider plugin — Hermes's pluggable interface for agent memory. Same automatic retrieval and extraction, delivered through Hermes's memory lifecycle instead of Claude Code hooks.

Install:

# Copy or symlink the plugin into Hermes's plugin directory
cp -r /path/to/ClawMem/src/hermes /path/to/hermes-agent/plugins/memory/clawmem

# Or symlink for development
ln -s /path/to/ClawMem/src/hermes /path/to/hermes-agent/plugins/memory/clawmem

Configure in your Hermes profile's .env or environment:

CLAWMEM_BIN=/path/to/clawmem          # Path to clawmem binary (or ensure it's on PATH)
CLAWMEM_SERVE_PORT=7438                # REST API port (default: 7438)
CLAWMEM_SERVE_MODE=external            # "external" (you run clawmem serve) or "managed" (plugin manages it)
CLAWMEM_PROFILE=balanced               # speed | balanced | deep

Then set memory.provider: clawmem in your Hermes config.yaml, or run hermes memory setup to configure interactively.

What the plugin provides:

• prefetch() — prompt-aware retrieval via context-surfacing hook (automatic every turn)
• on_session_end() — decision extraction, handoff generation, feedback loop (parallel)
• on_pre_compress() — pre-compaction state preservation
• session-bootstrap — session registration + first-turn context injection
• 5 agent tools — clawmem_retrieve, clawmem_get, clawmem_session_log, clawmem_timeline, clawmem_similar
• Plugin-managed transcript — maintains its own JSONL transcript for ClawMem hooks

Requirements: clawmem binary on PATH and clawmem serve running (external mode) or the plugin starts it automatically (managed mode). Python 3.10+. No pip dependencies beyond Hermes itself (uses urllib for REST calls, httpx optional for better performance).

Alternative: Hermes also has built-in MCP client support. You can add ClawMem as an MCP server in Hermes's config.yaml under mcp_servers for tool-only access. But this misses the lifecycle hooks (prefetch, session_end, pre_compress), so the native plugin is recommended.

See Hermes plugin guide for architecture details, lifecycle mapping, and troubleshooting.

Multi-Framework Operation

All three integrations share the same SQLite vault by default. Claude Code, OpenClaw, and Hermes can run simultaneously — decisions captured in one runtime are immediately available in the others, giving agents persistent shared memory across sessions and platforms. WAL mode + busy_timeout handles concurrent access.

Multi-Vault (Optional)

By default, ClawMem uses a single vault at ~/.cache/clawmem/index.sqlite. For users who want separate memory domains (e.g., work vs personal, or isolated vaults per project), ClawMem supports named vaults.

Configure in ~/.config/clawmem/config.yaml:

vaults:
  work: ~/.cache/clawmem/work.sqlite
  personal: ~/.cache/clawmem/personal.sqlite

Or via environment variable:

export CLAWMEM_VAULTS='{"work":"~/.cache/clawmem/work.sqlite","personal":"~/.cache/clawmem/personal.sqlite"}'

Using vaults with MCP tools:

All retrieval tools (memory_retrieve, query, search, vsearch, intent_search) accept an optional vault parameter. Omit it to use the default vault.

# Search the default vault (no vault param needed)
query("authentication flow")

# Search a named vault
query("project timeline", vault="work")

# List configured vaults
list_vaults()

# Sync content into a vault
vault_sync(vault="work", content_root="~/work/docs")

Single-vault users: No action needed. Everything works without configuration. The vault parameter is always optional and ignored when no vaults are configured.

GPU Services

ClawMem uses three llama-server (llama.cpp) instances for neural inference. All three have in-process fallbacks via node-llama-cpp (auto-downloads on first use), so ClawMem works without a dedicated GPU. node-llama-cpp auto-detects the best available backend — Metal on Apple Silicon, Vulkan where available, CPU as last resort. With GPU acceleration (Metal/Vulkan), in-process inference is fast for these small models (0.3B–1.7B); on CPU-only systems it is significantly slower. For production use, run the servers via systemd services to prevent silent fallback.

GPU with VRAM to spare (12GB+, recommended): ZeroEntropy's distillation-paired stack delivers best retrieval quality — total ~10GB VRAM.

Important: zembed-1 and zerank-2 use non-causal attention — -ub must equal -b on llama-server (e.g. -b 2048 -ub 2048). See Reranker Server for details.

License: zembed-1 and zerank-2 are released under CC-BY-NC-4.0 — non-commercial only. The QMD native models below have no such restriction.

No dedicated GPU / GPU without VRAM to spare: The QMD native combo — total ~4GB VRAM, also runs via node-llama-cpp (Metal on Apple Silicon, Vulkan where available, CPU as last resort). Fast with GPU acceleration; significantly slower on CPU-only.

The bin/clawmem wrapper defaults to localhost:8088/8089/8090. If a server is unreachable (transport error like ECONNREFUSED/ETIMEDOUT), ClawMem sets a 60-second cooldown and falls back to in-process inference via node-llama-cpp (auto-downloads the QMD native models on first use, uses Metal/Vulkan/CPU depending on hardware). HTTP errors (400/500) and user-cancelled requests do not trigger cooldown — the remote server is retried normally on the next call. With GPU acceleration the fallback is fast; on CPU-only it is significantly slower. ClawMem always works either way, but if you're running dedicated GPU servers, use systemd services to ensure they stay up.

To prevent fallback and fail fast instead, set CLAWMEM_NO_LOCAL_MODELS=true.

Remote GPU (optional)

If your GPU lives on a separate machine, point the env vars at it:

export CLAWMEM_EMBED_URL=http://gpu-host:8088
export CLAWMEM_LLM_URL=http://gpu-host:8089
export CLAWMEM_RERANK_URL=http://gpu-host:8090

For remote setups, set CLAWMEM_NO_LOCAL_MODELS=true to prevent node-llama-cpp from auto-downloading multi-GB model files if a server is unreachable.

No Dedicated GPU (in-process inference)

All three QMD native models run locally without a dedicated GPU. node-llama-cpp auto-downloads them on first use (~300MB embedding + ~1.1GB LLM + ~600MB reranker) and auto-detects the best backend — Metal on Apple Silicon (fast, uses integrated GPU), Vulkan where available (fast, uses discrete or integrated GPU), or CPU as last resort (significantly slower). With Metal or Vulkan, in-process inference handles these small models well; CPU-only is functional but noticeably slower.

Alternatively, use a cloud embedding provider if you prefer not to run models locally.

Embedding

ClawMem calls the OpenAI-compatible /v1/embeddings endpoint for all embedding operations. This works with local llama-server instances and cloud providers alike.

Option A: GPU with VRAM to spare (recommended)

Use zembed-1-Q4_K_M — SOTA retrieval quality, distilled from zerank-2 via ZeroEntropy's zELO methodology. CC-BY-NC-4.0 — non-commercial only.

• Size: 2.4GB, Dimensions: 2560, VRAM: ~4.4GB, Context: 32K tokens

wget https://huggingface.co/Abhiray/zembed-1-Q4_K_M-GGUF/resolve/main/zembed-1-Q4_K_M.gguf

# -ub must match -b for non-causal attention
llama-server -m zembed-1-Q4_K_M.gguf \
  --embeddings --port 8088 --host 0.0.0.0 \
  -ngl 99 -c 8192 -b 2048 -ub 2048

Option B: No GPU / GPU without VRAM to spare

Use EmbeddingGemma-300M-Q8_0 — the QMD native embedding model. Only 300MB, runs on CPU or any GPU.

• Size: 314MB, Dimensions: 768, VRAM: ~400MB (or CPU), Context: 2048 tokens

wget https://huggingface.co/ggml-org/embeddinggemma-300M-GGUF/resolve/main/embeddinggemma-300M-Q8_0.gguf

# On GPU (add -ngl 99):
llama-server -m embeddinggemma-300M-Q8_0.gguf \
  --embeddings --port 8088 --host 0.0.0.0 \
  -ngl 99 -c 2048 --batch-size 2048

# On CPU (omit -ngl):
llama-server -m embeddinggemma-300M-Q8_0.gguf \
  --embeddings --port 8088 --host 0.0.0.0 \
  -c 2048 --batch-size 2048

For multilingual corpora, the SOTA zembed-1 (Option A) supports multilingual out of the box. For a lightweight alternative: granite-embedding-278m-multilingual-Q6_K (314MB, set CLAWMEM_EMBED_MAX_CHARS=1100 due to 512-token context).

Option C: Cloud Embedding API

Alternatively, use a cloud embedding provider instead of running a local server. Any provider with an OpenAI-compatible /v1/embeddings endpoint works.

Configuration: Copy .env.example to .env and set your provider credentials:

cp .env.example .env
# Edit .env:
CLAWMEM_EMBED_URL=https://api.jina.ai
CLAWMEM_EMBED_API_KEY=jina_your-key-here
CLAWMEM_EMBED_MODEL=jina-embeddings-v5-text-small

Or export them in your shell. Precedence: shell environment > .env file > bin/clawmem wrapper defaults.

Cloud mode auto-detects your provider from the URL and sends the right parameters (Jina task, Voyage/Cohere input_type, OpenAI dimensions). Batch embedding (50 fragments/request), server-side truncation, adaptive TPM-aware pacing, and retry with jitter are all handled automatically. Set CLAWMEM_EMBED_TPM_LIMIT to match your provider tier (default: 100000). See docs/guides/cloud-embedding.md for full details.

Note: Cloud providers handle their own context window limits — ClawMem skips client-side truncation when an API key is set. Local llama-server truncates at CLAWMEM_EMBED_MAX_CHARS (default: 6000 chars).

Verify and embed

# Verify endpoint is reachable
curl $CLAWMEM_EMBED_URL/v1/embeddings \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $CLAWMEM_EMBED_API_KEY" \
  -d "{\"input\":\"test\",\"model\":\"$CLAWMEM_EMBED_MODEL\"}"

# Embed your vault
./bin/clawmem embed

LLM Server

Intent classification, query expansion, and A-MEM extraction use qmd-query-expansion-1.7B — a Qwen3-1.7B finetuned by QMD specifically for generating search expansion terms (hyde, lexical, and vector variants). ~1.1GB at q4_k_m quantization, served via llama-server on port 8089.

Without a server: If CLAWMEM_LLM_URL is unset, node-llama-cpp auto-downloads the model on first use.

Performance (RTX 3090):

• Intent classification: 27ms
• Query expansion: 333 tok/s
• VRAM: ~2.2-2.8GB depending on quantization

Qwen3 /no_think flag: Qwen3 uses thinking tokens by default. ClawMem appends /no_think to all prompts automatically to get structured output in the content field.

Intent classification: Uses a dual-path approach:

1. Heuristic regex classifier (instant) — handles strong signals (why/when/who keywords) with 0.8+ confidence
2. LLM refinement (27ms on GPU) — only for ambiguous queries below 0.8 confidence

Server setup:

# Download the finetuned model
wget https://huggingface.co/tobil/qmd-query-expansion-1.7B-gguf/resolve/main/qmd-query-expansion-1.7B-q4_k_m.gguf

# Start llama-server for LLM inference
llama-server -m qmd-query-expansion-1.7B-q4_k_m.gguf \
  --port 8089 --host 0.0.0.0 \
  -ngl 99 -c 4096 --batch-size 512

Reranker Server

Cross-encoder reranking for query and intent_search pipelines on port 8090. ClawMem calls the /v1/rerank endpoint (or falls back to scoring via /v1/completions for compatible servers).

Scores each candidate against the original query (cross-encoder architecture). query pipeline: 4000 char context per doc (deep reranking); intent_search: 200 char context per doc (fast reranking).

GPU with VRAM to spare (recommended): zerank-2-Q4_K_M (2.4GB, ~3.3GB VRAM). Outperforms Cohere rerank-3.5 and Gemini 2.5 Flash. Optimal pairing with zembed-1 (same distillation architecture via zELO). CC-BY-NC-4.0 — non-commercial only.

wget https://huggingface.co/keisuke-miyako/zerank-2-gguf-q4_k_m/resolve/main/zerank-2-Q4_k_m.gguf

# -ub must match -b for non-causal attention
llama-server -m zerank-2-Q4_K_M.gguf \
  --reranking --port 8090 --host 0.0.0.0 \
  -ngl 99 -c 2048 -b 2048 -ub 2048

CPU / GPU without VRAM to spare: qwen3-reranker-0.6B-Q8_0 (~600MB, ~1.3GB VRAM). The QMD native reranker — auto-downloaded by node-llama-cpp if no server is running.

wget https://huggingface.co/ggml-org/Qwen3-Reranker-0.6B-Q8_0-GGUF/resolve/main/Qwen3-Reranker-0.6B-Q8_0.gguf

llama-server -m Qwen3-Reranker-0.6B-Q8_0.gguf \
  --reranking --port 8090 --host 0.0.0.0 \
  -ngl 99 -c 2048 --batch-size 512

Note: zerank-2 and zembed-1 use non-causal attention — -ub (ubatch) must equal -b (batch). Omitting -ub or setting it lower causes assertion crashes. qwen3-reranker-0.6B does not have this requirement. See llama.cpp#12836.

MCP Server

ClawMem exposes 31 MCP tools via the Model Context Protocol and an optional HTTP REST API. Any MCP-compatible client or HTTP client can use it.

Claude Code (automatic):

./bin/clawmem setup mcp   # Registers in ~/.claude.json

Manual (any MCP client):

Add to your MCP config (e.g. ~/.claude.json, claude_desktop_config.json, or your client's equivalent):

{
  "mcpServers": {
    "clawmem": {
      "command": "/absolute/path/to/clawmem/bin/clawmem",
      "args": ["mcp"]
    }
  }
}

The server runs via stdio — no network port needed. The bin/clawmem wrapper sets the GPU endpoint env vars automatically.

Verify: After registering, your client should see tools including memory_retrieve, search, vsearch, query, query_plan, intent_search, timeline, etc.

HTTP REST API (optional)

For web dashboards, non-MCP agents, cross-machine access, or programmatic use:

./bin/clawmem serve                          # localhost:7438, no auth
./bin/clawmem serve --port 8080              # custom port
CLAWMEM_API_TOKEN=secret ./bin/clawmem serve # with bearer token auth

Endpoints:

Auth: Set CLAWMEM_API_TOKEN env var to require Authorization: Bearer <token> on all requests. If unset, access is open (localhost-only by default). See .env.example.

Search example:

curl -X POST http://localhost:7438/search \
  -H 'Content-Type: application/json' \
  -d '{"query": "authentication decisions", "mode": "hybrid", "compact": true}'

Verify Installation

./bin/clawmem doctor   # Full health check
./bin/clawmem status   # Quick index status
bun test               # Run test suite

Agent Instructions

ClawMem ships three instruction files and an optional maintenance agent:

Working in the ClawMem repo: No action needed — CLAUDE.md loads automatically.

Using ClawMem from other projects: Your agent needs instructions on how to use ClawMem's hooks and MCP tools. Two options:

Option A: Copy instructions into your project

Copy the contents of CLAUDE.md (or the relevant sections) into your project's own CLAUDE.md or AGENTS.md. Simple but requires manual updates when ClawMem changes.

Option B: Add a trigger block (recommended)

Add this minimal trigger block to your global ~/.claude/CLAUDE.md. It gives the agent routing rules always loaded, and tells it how to find the full reference (SKILL.md) shipped with your installation when deeper guidance is needed:

## ClawMem

Architecture: hooks (automatic, ~90%) + MCP tools (explicit, ~10%).

Vault: `~/.cache/clawmem/index.sqlite` | Config: `~/.config/clawmem/config.yaml`

### Escalation Gate (3 rules — ONLY escalate to MCP tools when one fires)

1. **Low-specificity injection** — `<vault-context>` is empty or lacks the specific fact needed
2. **Cross-session question** — "why did we decide X", "what changed since last time"
3. **Pre-irreversible check** — before destructive or hard-to-reverse changes

### Tool Routing (once escalated)

**Preferred:** `memory_retrieve(query)` — auto-classifies and routes to the optimal backend.

**Direct routing** (when calling specific tools):

    "why did we decide X"         → intent_search(query)          NOT query()
    "what happened last session"  → session_log()                 NOT query()
    "what else relates to X"      → find_similar(file)            NOT query()
    Complex multi-topic           → query_plan(query)             NOT query()
    General recall                → query(query, compact=true)
    Keyword spot check            → search(query, compact=true)
    Conceptual/fuzzy              → vsearch(query, compact=true)
    Full content                  → multi_get("path1,path2")
    Lifecycle health              → lifecycle_status()
    Stale sweep                   → lifecycle_sweep(dry_run=true)
    Restore archived              → lifecycle_restore(query)

ALWAYS `compact=true` first → review → `multi_get` for full content.

### Proactive Use (no escalation gate needed)

- User says "remember this" / critical decision made → `memory_pin(query)` immediately
- User corrects a misconception → `memory_pin(query)` the correction
- `<vault-context>` surfaces irrelevant/noisy content → `memory_snooze(query, until)` for 30 days
- Need to correct a memory → `memory_forget(query)`
- After bulk ingestion → `build_graphs`

### Anti-Patterns

- Do NOT use `query()` for everything — match query type to tool, or use `memory_retrieve`
- Do NOT call query/intent_search every turn — 3 rules above are the only gates
- Do NOT re-search what's already in `<vault-context>`
- Do NOT pin everything — pin is for persistent high-priority items, not routine decisions
- Do NOT forget memories to "clean up" — let confidence decay handle it
- Do NOT wait for curator to pin decisions — pin immediately when critical

For detailed operational guidance (query optimization, troubleshooting, collection setup, embedding workflow, graph building, curator), find and read the shipped SKILL.md:
  Bash: CLAWMEM_ROOT=$(cd "$(dirname "$(which clawmem)")/.." && pwd) && echo "$CLAWMEM_ROOT/SKILL.md"
  Then: Read the file at that path.

This gives your agent the 3-rule gate, tool routing, and proactive behaviors always loaded. When it needs deeper guidance, it locates and reads the full SKILL.md reference shipped with your installation — no symlinks or skill registration required.

CLI Reference

clawmem init                                    Create DB + config
clawmem bootstrap <vault> [--name N] [--skip-embed]  One-command setup
clawmem collection add <path> --name <name>     Add a collection
clawmem collection list                         List collections
clawmem collection remove <name>                Remove a collection

clawmem update [--pull] [--embed]               Incremental re-scan
clawmem mine <dir> [-c name] [--embed] [--synthesize]  Import conversation exports (--synthesize runs post-import LLM fact extraction, v0.7.2)
clawmem embed [-f]                              Generate fragment embeddings
clawmem reindex [--force]                       Full re-index
clawmem watch                                   File watcher daemon

clawmem search <query> [-n N] [--json]          BM25 keyword search
clawmem vsearch <query> [-n N] [--json]         Vector semantic search
clawmem query <query> [-n N] [--json]           Full hybrid pipeline

clawmem profile                                 Show user profile
clawmem profile rebuild                         Force profile rebuild
clawmem update-context                          Regenerate per-folder CLAUDE.md

clawmem list [-n/--limit N] [-c col] [--json]   Browse recent documents
clawmem budget [--session ID]                   Token utilization
clawmem log [--last N]                          Session history
clawmem hook <name>                             Manual hook trigger
clawmem surface --context --stdin               IO6: pre-prompt context injection
clawmem surface --bootstrap --stdin             IO6: per-session bootstrap injection

clawmem reflect [N]                             Cross-session reflection (last N days, default 14)
clawmem consolidate [--dry-run] [N]             Find and archive duplicate low-confidence docs

clawmem install-service [--enable] [--remove]   Systemd watcher service
clawmem setup hooks [--remove]                  Install/remove Claude Code hooks
clawmem setup mcp [--remove]                    Register/remove MCP server
clawmem setup curator [--remove]                Install/remove curator maintenance agent
clawmem mcp                                     Start stdio MCP server
clawmem serve [--port 7438] [--host 127.0.0.1]  Start HTTP REST API server
clawmem path                                    Print database path
clawmem doctor                                  Full health check
clawmem status                                  Quick index status

MCP Tools (31)

Registered by clawmem setup mcp. Available to any MCP-compatible client.

Core Search & Retrieval

Intent-Aware Search

intent_search pipeline: Query → Intent Classification → BM25 + Vector → Intent-Weighted RRF → Graph Expansion (WHY/ENTITY intents) → Cross-Encoder Reranking → Composite Scoring

query_plan pipeline: Query → LLM decomposition into 2-4 typed clauses → Parallel execution (BM25/vector/graph per clause) → RRF merge across clauses → Composite scoring. Falls back to single-query for simple inputs.

Multi-Graph & Causal

Beads Integration

Vault Management

Agent Diary

Memory Management & Lifecycle

Compact Mode

search, vsearch, and query accept compact: true to return { id, path, title, score, snippet, content_type, fragment } instead of full content. Saves ~5x tokens for initial filtering.

Hooks (Claude Code Integration)

Hooks installed by clawmem setup hooks:

| Hook | Event | Wha

README truncated. Continue reading on GitHub