MemoryMesh/.env.example at main · staavanothanh/MemoryMesh · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
# ============================================
# MemoryMesh Configuration
# ============================================
# Copy this file to .env and adjust for your environment.
# All variables are optional — defaults are shown below.

# --- LLM Router ---
# Point to any OpenAI-compatible endpoint (Ollama, vLLM, OpenAI, etc.)
ROUTER_URL=http://127.0.0.1:20128/v1
DEFAULT_MODEL=your-model
FALLBACK_MODEL=your-fallback-model
# Background LLM pool — cascade qua free models cho compaction/snapshot/facts
# Thứ tự ưu tiên: Gemini (context 1M) → DeepSeek (JSON clean) → Llama (reasoning) → Qwen (code)
BACKGROUND_MODEL_POOL=openrouter/google/gemini-2.5-flash,openrouter/deepseek/deepseek-v4-flash,openrouter/qwen/qwen-2.5-7b-instruct,openrouter/meta-llama/llama-3.1-8b-instruct
ROUTER_TIMEOUT=30
ROUTER_RETRIES=3

# --- Embedding ---
# sentence-transformers model (downloaded on first use)
EMBEDDING_MODEL=paraphrase-multilingual-MiniLM-L12-v2

# --- Storage ---
VEC_DB_PATH=./db/memory.db
SESSION_DB_PATH=./db/sessions.db
INSTINCT_DB_PATH=./db/instincts.db
VEC_AUTO_MIGRATE=false

# --- Background Tasks ---
# Set to false to disable automatic atomic fact extraction (saves token cost)
AUTO_EXTRACT_FACTS=true

# --- User ---
DEFAULT_USER_ID=your_user_id
# --- Transport ---
# stdio (default) for CLI agents, sse for HTTP clients
MCP_TRANSPORT=stdio
MCP_PORT=8090

# --- Session Start Instructions ---
# Priority-ordered list of instruction files to auto-load (comma-separated).
# Scans workspace root for these filenames — first match wins.
# Supports: opencode.md, CLAUDE.md, .cursorrules, .windsurfrules, memorymesh.md, ...
SESSION_INSTRUCTION_FILES=opencode.md,CLAUDE.md,.cursorrules,memorymesh.md
# Directory for instruction fragments (all .md/.txt files merged, sorted)
SESSION_INSTRUCTIONS_DIR=.memorymesh/instructions/
# Global user-level instruction file (supports ~ expansion)
SESSION_GLOBAL_INSTRUCTION=~/.config/memorymesh/session.md
# Max file size in bytes for instruction/doc files
SESSION_INSTRUCTION_MAX_FILE_SIZE=51200
# --- Session Docs Sync ---
# Auto-sync project docs (README, Makefile, etc.) into memory on session start
SESSION_DOCS_SYNC_ENABLED=true
# Comma-separated list of doc files to sync
SESSION_DOCS_SYNC_FILES=README.md,CONTRIBUTING.md,CHANGELOG.md,Makefile

# --- Logging ---
LOG_LEVEL=INFO