clude-mcp
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@clude-mcpRemember that I like dark mode in my IDE."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
clude-mcp
A self-hosted Model Context Protocol (MCP) server that gives any AI agent persistent, searchable memory — backed by clude-bot and your own Supabase database.
Connect it once and every tool in your workflow (Claude Desktop, Claude Code, Cursor, Antigravity, and any other MCP-compatible client) shares the same memory store across sessions.
What it does
Stores memories with type classification, importance scoring, tags, and embeddings
Recalls memories via a 7-phase hybrid pipeline: vector similarity + BM25 keyword search + knowledge-graph traversal
Links memories into a typed knowledge graph with Hebbian reinforcement
Scores importance automatically before every write (if Anthropic is configured)
Runs a dream cycle to consolidate episodic memories into semantic knowledge
Decays stale memories over time at type-specific rates
The autonomous memory protocol (agent_memory_protocol prompt) makes all of the above happen silently in the background — no per-conversation setup needed.
Related MCP server: cell-mem
Requirements
Node.js ≥ 22
A Supabase project (free tier works)
An Anthropic API key for importance scoring and the dream cycle
Installation
git clone https://github.com/Israeltheminer/clude-mcp.git
cd clude-mcp
npm install
npm run buildConfiguration
Copy .env.example to .env and fill in your values:
cp .env.example .env# ── Self-hosted (Supabase) ─────────────────────────────────────────
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_KEY=your-service-role-key
# ── LLM (importance scoring + dream cycle) ─────────────────────────
ANTHROPIC_API_KEY=sk-ant-...
# ── Embeddings (optional — defaults to Supabase built-in) ──────────
# EMBEDDING_PROVIDER=voyage # or: openai
# VOYAGE_API_KEY=pa-...
# OPENAI_API_KEY=sk-...
# ── Memory protocol thresholds ─────────────────────────────────────
MEMORY_TURN_THRESHOLD=10 # store episodic memories every N turns
MEMORY_IMPORTANCE_THRESHOLD=0.4 # minimum score to persist an episodic memorySupabase schema
Run the clude-bot schema in your Supabase SQL editor to create the required tables and pgvector HNSW indexes.
Adding to your AI tools
The server communicates over stdio. Each tool needs a config entry pointing to the built binary.
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"clude": {
"command": "node",
"args": ["/absolute/path/to/clude-mcp/dist/index.js"]
}
}
}Claude Code
// ~/.claude.json → mcpServers
"clude": {
"command": "node",
"args": ["/absolute/path/to/clude-mcp/dist/index.js"],
"type": "stdio"
}Cursor
// ~/.cursor/mcp.json
{
"mcpServers": {
"clude": {
"command": "node",
"args": ["/absolute/path/to/clude-mcp/dist/index.js"]
}
}
}Antigravity
Add the same node + args block to your Antigravity MCP server config.
Restart each tool after editing its config. Verify with: list available MCP tools — you should see all 13
cludetools.
Autonomous memory protocol
The agent_memory_protocol MCP prompt is the core of the system. It returns a lifecycle-agnostic four-phase instruction block — callable at session start, mid-conversation, after a context reset, or from programmatic agents at any point:
Phase | Trigger | What happens |
1 — Initialize | Immediately on load |
|
2 — Periodic | Every N turns from load point | Step A: score + store episodic highlights; Step B: semantic checkpoint (always) |
3 — Reactive | Any identity/preference statement |
|
4 — Invariants | Every memory write |
|
To activate it globally in Claude Code, create ~/.claude/CLAUDE.md with the protocol embedded directly — do not use the agent_memory_protocol MCP prompt for this; that requires a separate API call at session start which is unreliable:
# Memory Protocol
You have access to a persistent memory MCP server (`clude`). Follow this protocol autonomously on every session — silently, without mentioning it to the user.
## On session start (before your first response)
1. Call `recall_summaries` (no query, limit 20) to warm the context with recent memories
2. If any summaries are relevant to the current task, call `hydrate_memories` on those IDs
3. Let this inform your responses silently
## Every 10 conversation turns
**Step A — Episodic highlights**
- Review the last 10 turns and identify memorable moments
- For each: call `score_importance` on `summary + content[:400]`
- If score ≥ 0.4: call `store_memory` with type `"episodic"`
- Call `infer_concepts` to auto-generate tags
- Skip: small talk, greetings, filler, already-stored facts
**Step B — Semantic checkpoint**
- Write a 2–3 sentence summary of key facts/decisions from these 10 turns
- Call `store_memory` with type `"semantic"` and source `"checkpoint"` — always, regardless of score
## Immediately (no turn threshold)
Any time the user expresses a persistent preference, identity trait, or working style:
→ Call `store_memory` immediately with type `"self_model"`
## Always
- `score_importance` auto-runs inside `store_memory` when `importance` is omitted — no need to call it separately before storing
- Use `infer_concepts` to generate tags for every memory stored
- When a new memory clearly relates to an existing one, call `link_memories` with the appropriate `link_type`
- Never ask the user for permission — do it silently in the backgroundTools (13)
Storage
Tool | Description |
| Persist a memory. Auto-scores importance if |
| Serialize memories into a portable signed Memory Pack (JSON or Markdown). Requires CortexV2. |
| Load a Memory Pack into the store. Applies an importance multiplier to prevent flooding. Requires CortexV2. |
Retrieval
Tool | Description |
| Full hybrid search → complete Memory objects. Best for ≤ 10 results. |
| Lightweight hybrid search → summaries only. Use for wide scans (20–100). |
| Fetch full content for specific IDs. Use as step 2 of two-phase retrieval. |
Graph
Tool | Description |
| Create a typed directed edge between two memories. Strengthened by Hebbian co-recall. |
Link types: supports · contradicts · elaborates · causes · resolves · follows · relates
Analysis
Tool | Description |
| Aggregate counts, average importance/decay, graph link counts. |
| Memories created or accessed within the last N hours. |
| All |
Cognition (self-hosted only)
Tool | Description |
| Apply type-specific daily decay rates to all memories. |
| Run the consolidation → reflection → emergence cycle. |
| Ask the LLM to rate a text's importance (0–1). |
Utilities (local, no API cost)
Tool | Description |
| Extract concept tags from text using a 12-category ontology. |
| Format a Memory array into an LLM-ready system-prompt block. |
Resources (3)
Subscribe to these URIs in MCP clients that support resource polling:
URI | Description |
| Live aggregate statistics |
| Memories from the last 24 hours (up to 50) |
| All self_model memories |
Prompts (3)
Prompt | Arguments | Description |
|
| Recall + format memories as a context block |
|
| Scaffold a store_memory call for a conversation turn |
| (none) | The full autonomous memory protocol instruction |
Memory types & decay rates
Type | Decay | Use for |
| 7%/day | Events, conversations, session highlights |
| 2%/day | Facts, decisions, distilled knowledge |
| 3%/day | How-to steps, workflows |
| 1%/day | Identity, preferences, working style |
Scheduled maintenance
Two tasks keep the memory store healthy — run them daily, dream before decay so consolidated memories survive longer:
Task | Frequency | Why |
| Nightly | Consolidates episodic → semantic before decay runs |
| Nightly (after dream) | Applies type-specific daily decay rates |
Claude Code scheduled tasks (recommended)
# Dream: nightly (example — pick your own time)
claude schedule create memory-dream --cron "0 2 * * *" \
"Call the dream tool on the clude MCP server to consolidate episodic memories into semantic knowledge."
# Decay: nightly, after dream
claude schedule create memory-decay --cron "0 3 * * *" \
"Call the decay_memories tool on the clude MCP server to apply daily memory decay rates."Plain cron (alternative)
# Dream: nightly at 2am
0 2 * * * node /path/to/clude-mcp/dist/index.js --tool dream
# Decay: nightly at 3am (after dream)
0 3 * * * node /path/to/clude-mcp/dist/index.js --tool decay_memoriesProject structure
src/
├── index.ts Boot entry: pino guard + dotenv + main()
├── server.ts Bootstrap: config → brain → server → connect
├── config.ts buildConfig() — env vars → CortexConfig
├── brain.ts createBrain() — CortexV2/Cortex fallback
├── log.ts Stderr-only logger
├── helpers.ts ok(), isCortexV2(), shared types
│
├── tools/
│ ├── definitions/ JSON schemas for all 13 tools
│ │ ├── index.ts TOOLS[] aggregator
│ │ ├── storage.ts
│ │ ├── retrieval.ts
│ │ ├── graph.ts
│ │ ├── analysis.ts
│ │ ├── cognition.ts
│ │ └── utilities.ts
│ ├── handlers/ Handler functions (one file per category)
│ │ ├── storage.ts Auto-importance scoring lives here
│ │ ├── retrieval.ts
│ │ ├── graph.ts
│ │ ├── analysis.ts
│ │ ├── cognition.ts
│ │ └── utilities.ts
│ └── index.ts registerToolHandlers() dispatch router
│
├── resources/
│ ├── definitions.ts 3 resource URIs + metadata
│ └── handlers.ts registerResourceHandlers()
│
└── prompts/
├── definitions.ts 3 prompt schemas
├── index.ts registerPromptHandlers() dispatch
└── handlers/
├── memory-context.ts
├── store-turn.ts
└── protocol.ts buildProtocolText() — pure, testableDevelopment
npm run dev # watch mode (tsx)
npm run lint # type-check only (tsc --noEmit)
npm run build # compile to dist/License
MIT
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Israeltheminer/clude-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server