MCP Memory Service

Overview

Decisions, fixes and notes your agents stored, searchable by meaning

— Total Memories

— This Week

— Tags

Recent Memories

Quick Actions

Checking...

Search Filters

Live Search mode

Search Results

0 results

Document Ingestion

Drag & drop files here

Supported formats: PDF, TXT, MD, JSON

Tags (comma-separated) Tags will be applied to all files. Use spaces or commas as separators.

Processing Mode

All selected files will be processed together with the same tags.

Chunk Size: 1000 chars

Overlap: 200 chars

Chunking Configuration Guide

Note: Actual chunk sizes may vary as the system respects paragraph boundaries to maintain semantic coherence.

Default (1000 chars, 200 overlap) Recommended

Best for: Technical documentation, reference manuals, knowledge bases

Why: Paragraph-aware chunking preserves complete thoughts and context

Smaller Chunks (500 chars, 100 overlap)

Best for: Dense technical docs, code documentation, API references

Trade-off: More granular retrieval but may split paragraphs more aggressively

Larger Chunks (2000 chars, 400 overlap)

Best for: Narrative documents, articles, blogs, long-form content

Trade-off: Better context but less precise retrieval

Tips:

Overlap helps maintain context across chunk boundaries
Higher overlap = better continuity but more redundancy
Test different settings on a sample document to find optimal configuration
Chunks preserve complete sentences and paragraphs when possible

Chunk Overlap Explained

What is overlap? Overlap is the number of characters that are duplicated between consecutive chunks. This helps maintain context across chunk boundaries.

Visual Example:

Chunk 1 unique content... shared overlap region

shared overlap region Chunk 2 unique content...

No Overlap (0 chars)

Best for: Maximum storage efficiency, no redundancy needed

Trade-off: Context may be lost at chunk boundaries

Medium Overlap (200 chars) Recommended

Best for: Most documents - balances context and efficiency

Why: Preserves 1-2 sentences of context across boundaries

High Overlap (400+ chars)

Best for: Complex technical content requiring maximum context

Trade-off: More storage and processing, higher redundancy

Guidelines:

Rule of thumb: Overlap should be 15-25% of chunk size
Small chunks (500) → Use 100-150 overlap
Medium chunks (1000) → Use 200-250 overlap
Large chunks (2000) → Use 400-500 overlap
Higher overlap helps with search accuracy but increases storage
Zero overlap is fine for well-structured documents with clear sections

Processing Mode Options

When uploading multiple files, choose how they should be processed. Both modes apply the same tags to all files.

Batch Processing Default

What it does: Uploads all files together as one operation

Best for: Similar files, fast bulk processing, when you want files grouped together

Pros: Faster, simpler, single progress indicator

Cons: One file failure can affect the whole batch

Individual Processing

What it does: Uploads each file separately with individual API calls

Best for: Mixed file types, when you want error isolation, or need individual progress tracking

Pros: Better error handling, individual progress, more robust

Cons: Slightly slower due to sequential processing

When to choose each mode:

Batch mode: When files are similar and you want them processed quickly together
Individual mode: When files might have different processing requirements or you want to ensure all files get processed even if some fail
Single files: Always processed individually (no choice needed)
Both modes apply the same tags to all files

Memory Type

Upload History

Loading upload history...

Search Ingested Content

Search within your uploaded documents to verify content is indexed

Browse by Tags

Explore your memories organized by tags

Bulk Operations

Delete by Tag

Remove all memories with a specific tag

Cleanup Duplicates

Remove duplicate memories based on content

Delete by Date

Remove memories older than a specific date

Delete Untagged

Remove memories without any tags (0 found)

Tag Management

Loading tag statistics...

System Maintenance

Database Optimization

Optimize database performance

Rebuild Search Index

Rebuild search indexes for better performance

Key Metrics

Total Memories

This Week

Unique Tags

Database Size

Trends & Charts

Memory Growth Over Time

Loading chart...

Tag Usage Distribution

Loading chart...

Memory Types Distribution

Loading chart...

Activity Heatmap

Loading heatmap...

Knowledge Graph

Explore typed relationships between your memories

Relationship Types

Loading chart...

Interactive Graph Visualization

Nodes: Min Connections: Min Degree: 1

Loading graph...

Legend

Node Colors:

Observation Decision Learning Error Other

Edge Types:

causes fixes related

Detailed Reports

Top Tags

Recent Activity

Storage Report

Memory Quality Analytics

Quality score of each memory and how the scores are distributed

Total Memories

High Quality (≥0.7)

Medium (0.5-0.7)

Low (<0.5)

Average Score

Quality Score Distribution

Scoring Provider Breakdown

Top Quality Memories

Memories for Improvement

API Documentation

REST endpoints of this server

Interactive Swagger UI ReDoc Documentation

Memory Management

POST /api/memories Store a new memory with automatic embedding generation

GET /api/memories List all memories with pagination support

GET /api/memories/{content_hash} Retrieve a specific memory by content hash

DELETE /api/memories/{content_hash} Delete a memory and its embeddings

Search Operations

POST /api/search Semantic similarity search using embeddings

POST /api/search/by-tag Search memories by tags (AND/OR logic)

POST /api/search/by-time Natural language time-based queries

GET /api/search/similar/{content_hash} Find memories similar to a specific one

Real-time Events

GET /api/events Subscribe to real-time memory events stream

GET /api/events/stats View SSE connection statistics

Health & Status

GET /api/health Quick health check endpoint

GET /api/health/detailed Detailed health with database statistics

Overview

Recent Memories

Quick Actions

Search Results

Document Ingestion

Drag & drop files here

Upload History

Search Ingested Content

Search Results

Browse by Tags

Memories tagged with:

Bulk Operations

Delete by Tag

Cleanup Duplicates

Delete by Date

Delete Untagged

Tag Management

System Maintenance

Database Optimization

Rebuild Search Index

Key Metrics

Trends & Charts

Memory Growth Over Time

Tag Usage Distribution

Memory Types Distribution

Activity Heatmap

Knowledge Graph

Relationship Types

Interactive Graph Visualization

Legend

Detailed Reports

Top Tags

Recent Activity

Storage Report

Memory Quality Analytics

Quality Score Distribution

Scoring Provider Breakdown

Top Quality Memories

Memories for Improvement

API Documentation

Memory Management

Search Operations

Real-time Events

Health & Status

Memory Details

Interactive Graph Visualization

Add New Memory

Settings

Preferences

System Information

OAuth Status

Cloudflare Credentials

Sync Configuration

Quality System

Backup & Restore

Server Management

Backup History

Server Restarting...

Document Memory Chunks

Loading...

Authentication Required

API Key

OAuth