Elasticsearch Knowledge Graph
About
Elasticsearch-based knowledge graph that tracks access patterns to prioritize recent, important, and frequently accessed information with advanced search capabilities and complete CRUD operations.
Details
- Repository
- j3k0/mcp-brain-tools
Explore
- Spaced repetition freshness — each entity has a review interval that doubles on verification (capped at 365 days). Confidence labels (fresh/normal/aging/stale/archival) tell agents what to trust.
- Progressive search — queries return fresh results first, automatically widening to include older data only when needed.
- Observations as entities — each observation gets its own freshness lifecycle, so "build is broken" (1-day review) and "founded in 2015" (365-day review) age independently.
- Memory zones — isolate knowledge by project, team, or domain.
- AI-powered filtering — optional Groq integration scores search results by relevance.
- DRY by design — tool descriptions guide agents not to store what's already in code, git, or docs.
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
Elasticsearch Knowledge GraphCommand (node, npx, python, etc.)nodeArguments-
Argument 1
/path/to/mcp-brain-tools/dist/index.js
Environment-
ES_NODE
http://localhost:9200 -
GROQ_API_KEY
your-key-here
Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
-
Argument 1
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
- Node.js >= 18
- Docker (for Elasticsearch) or a remote Elasticsearch instance
npm install
npm run build
Add to your Claude Code, Claude Desktop, or other MCP client config:
{
"mcpServers": {
"memory": {
"command": "node",
"args": ["/path/to/mcp-brain-tools/dist/index.js"],
"env": {
"ES_NODE": "http://localhost:9200",
"GROQ_API_KEY": "your-key-here"
}
}
}
}
GROQ_API_KEY is optional — enables AI-powered search filtering and zone relevance scoring.
The memory hook runs on every user message and automatically injects relevant context — no agent cooperation needed.
Add to ~/.claude/settings.json:
{
"hooks": {
"UserPromptSubmit": [
{
"hooks": [
{
"type": "command",
"command": "node /path/to/mcp-brain-tools/dist/memory-hook.js"
}
]
}
]
}
}
The hook uses the same ES_NODE, AI_API_KEY/GROQ_API_KEY, AI_API_BASE, and AI_MODEL env vars (set them in the env block of your settings, or export them in your shell profile).
AI_API_BASE defaults to Groq's endpoint but accepts any OpenAI-compatible API URL.
create_entities
Create entities with optional observations and reviewInterval.
update_entities
Update existing entities.
delete_entities
Delete entities (with optional cascade).
add_observations
Add observations as separate entities with own freshness.
verify_entity
Confirm entity is still accurate, extend review interval.
search_nodes
Search with progressive freshness filtering.
open_nodes
Get specific entities by name with freshness metadata.
get_recent
Get recently accessed entities.
create_relations
Create relationships between entities.
delete_relations
Remove relationships.
inspect_knowledge_graph
AI-powered entity retrieval with tentative answers.
inspect_files
AI-powered file content inspection.
list_zones
List memory zones (with AI relevance scoring).
create_zone
Manage memory zones.
delete_zone
Manage memory zones.
copy_entities
Transfer entities between zones.
move_entities
Transfer entities between zones.
merge_zones
Merge zones with conflict resolution.
zone_stats
Get entity/relation counts for a zone.
mark_important
Boost entity relevance score.
get_time_utc
Get current UTC time.
An MCP server that gives AI agents persistent memory with built-in freshness tracking and spaced repetition. Backed by Elasticsearch.
Unlike simple key-value memory stores, mcp-brain-tools tracks how old each piece of knowledge is, flags what needs review, and lets agents verify information to keep it fresh — inspired by how spaced repetition helps humans retain knowledge.
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"elasticsearch knowledge graph": {
"env": {
"ES_NODE": "http://localhost:9200",
"GROQ_API_KEY": "your-key-here"
},
"args": [
"/path/to/mcp-brain-tools/dist/index.js"
],
"command": "node"
}
}
}Linux
{
"env": {
"ES_NODE": "http://localhost:9200",
"GROQ_API_KEY": "your-key-here"
},
"args": [
"/path/to/mcp-brain-tools/dist/index.js"
],
"command": "node"
}Macos
{
"env": {
"ES_NODE": "http://localhost:9200",
"GROQ_API_KEY": "your-key-here"
},
"args": [
"/path/to/mcp-brain-tools/dist/index.js"
],
"command": "node"
}Windows
{
"env": {
"ES_NODE": "http://localhost:9200",
"GROQ_API_KEY": "your-key-here"
},
"args": [
"/path/to/mcp-brain-tools/dist/index.js"
],
"command": "node"
}mcp-brain-tools
An MCP server that gives AI agents persistent memory with built-in freshness tracking and spaced repetition. Backed by Elasticsearch.
Unlike simple key-value memory stores, mcp-brain-tools tracks how old each piece of knowledge is, flags what needs review, and lets agents verify information to keep it fresh — inspired by how spaced repetition helps humans retain knowledge.
Features
- Spaced repetition freshness — each entity has a review interval that doubles on verification (capped at 365 days). Confidence labels (fresh/normal/aging/stale/archival) tell agents what to trust.
- Progressive search — queries return fresh results first, automatically widening to include older data only when needed.
- Observations as entities — each observation gets its own freshness lifecycle, so "build is broken" (1-day review) and "founded in 2015" (365-day review) age independently.
- Memory zones — isolate knowledge by project, team, or domain.
- AI-powered filtering — optional Groq integration scores search results by relevance.
- DRY by design — tool descriptions guide agents not to store what's already in code, git, or docs.
Setup
Prerequisites
- Node.js >= 18
- Docker (for Elasticsearch) or a remote Elasticsearch instance
Install and build
npm install
npm run build
Start Elasticsearch
npm run es:start
Or point to your own instance via ES_NODE environment variable.
Configure your MCP client
Add to your Claude Code, Claude Desktop, or other MCP client config:
{
"mcpServers": {
"memory": {
"command": "node",
"args": ["/path/to/mcp-brain-tools/dist/index.js"],
"env": {
"ES_NODE": "http://localhost:9200",
"GROQ_API_KEY": "your-key-here"
}
}
}
}
GROQ_API_KEY is optional — enables AI-powered search filtering and zone relevance scoring.
Install the auto-memory hook (Claude Code only)
The memory hook runs on every user message and automatically injects relevant context — no agent cooperation needed.
Add to ~/.claude/settings.json:
{
"hooks": {
"UserPromptSubmit": [
{
"hooks": [
{
"type": "command",
"command": "node /path/to/mcp-brain-tools/dist/memory-hook.js"
}
]
}
]
}
}
The hook uses the same ES_NODE, AI_API_KEY/GROQ_API_KEY, AI_API_BASE, and AI_MODEL env vars (set them in the env block of your settings, or export them in your shell profile).
AI_API_BASE defaults to Groq's endpoint but accepts any OpenAI-compatible API URL.
How it works
Entities and observations
Entities represent anything worth remembering — people, projects, decisions, facts. Each entity has:
- A name and type
- Spaced repetition fields: verifiedAt, reviewInterval, nextReviewAt
- A confidence label computed from freshness: 1 - (daysSinceVerified / reviewInterval)
Observations are stored as separate entities linked via is_observation_of relations. Each observation has its own review cadence:
```
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
