Skip to main content
Glama

Mnemex

.env.exampleβ€’2.08 kB
# Mnemex Configuration # Storage (Phase 2 - JSONL) # JSONL storage directory (contains memories.jsonl and relations.jsonl) MNEMEX_STORAGE_PATH=~/.config/mnemex/jsonl # Decay Model Selection # Options: power_law | exponential | two_component MNEMEX_DECAY_MODEL=power_law # Power-law parameters (default model) MNEMEX_PL_ALPHA=1.1 MNEMEX_PL_HALFLIFE_DAYS=3.0 # Exponential parameters (if MNEMEX_DECAY_MODEL=exponential) # MNEMEX_DECAY_LAMBDA=2.673e-6 # 3-day half-life # Two-component parameters (if MNEMEX_DECAY_MODEL=two_component) # MNEMEX_TC_LAMBDA_FAST=1.603e-5 # ~12h half-life # MNEMEX_TC_LAMBDA_SLOW=1.147e-6 # ~7d half-life # MNEMEX_TC_WEIGHT_FAST=0.7 # Common decay parameters MNEMEX_DECAY_BETA=0.6 # Thresholds # Minimum score before forgetting MNEMEX_FORGET_THRESHOLD=0.05 # Score threshold for automatic promotion MNEMEX_PROMOTE_THRESHOLD=0.65 # Use count threshold for promotion MNEMEX_PROMOTE_USE_COUNT=5 # Time window for use count (days) MNEMEX_PROMOTE_TIME_WINDOW=14 # Embeddings (optional) # Sentence transformer model for semantic search MNEMEX_EMBED_MODEL=all-MiniLM-L6-v2 # Enable embeddings (true/false) MNEMEX_ENABLE_EMBEDDINGS=false # Semantic Search Thresholds # High similarity = likely duplicate MNEMEX_SEMANTIC_HI=0.88 # Low similarity = likely distinct MNEMEX_SEMANTIC_LO=0.78 # Clustering # Cosine similarity threshold for linking in clusters MNEMEX_CLUSTER_LINK_THRESHOLD=0.83 # Maximum cluster size for LLM review MNEMEX_CLUSTER_MAX_SIZE=12 # Long-Term Memory (LTM) Integration # Path to Obsidian vault for LTM storage and search LTM_VAULT_PATH=~/Documents/Obsidian/Vault # Path to LTM index file LTM_INDEX_PATH=~/.config/mnemex/ltm-index.jsonl # Folder within vault for promoted memories LTM_PROMOTED_FOLDER=stm-promoted # Git Backup # Enable automatic git commits GIT_AUTO_COMMIT=true # Auto-commit interval in seconds (3600 = 1 hour) GIT_COMMIT_INTERVAL=3600 # Unified Search # Weight for STM results in unified search SEARCH_STM_WEIGHT=1.0 # Weight for LTM results in unified search SEARCH_LTM_WEIGHT=0.7 # Logging LOG_LEVEL=INFO

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/simplemindedbot/mnemex'

If you have feedback or need assistance with the MCP directory API, please join our Discord server