Configuration

Memex environment variables and options

Memex supports configuration via environment variables or config file. All settings have sensible defaults.

Priority: Environment variables > Config file > Defaults

Config File

Optional JSON config at ~/.vimo/memex/config.json:

{
  "server": {
    "port": 10013
  },
  "ollama": {
    "api": "http://localhost:11434",
    "embedding_model": "bge-m3",
    "chat_model": "qwen3:0.6b"
  },
  "features": {
    "enable_ai_chat": false
  },
  "compact": {
    "enabled": false,
    "fts_tokenizer": "trigram"
  }
}

Environment Variables

Variable	Default	Description
`PORT`	`10013`	HTTP server port
`VIMO_HOME`	`~/.vimo`	Base directory for all data
`CLAUDE_PROJECTS_PATH`	`~/.claude/projects`	Claude Code sessions
`CODEX_PATH`	`~/.codex`	Codex CLI sessions
`OPENCODE_PATH`	`~/.local/share/opencode`	OpenCode sessions
`GEMINI_TMP_PATH`	`~/.gemini/tmp`	Gemini CLI sessions
`OLLAMA_API`	`http://localhost:11434`	Ollama API endpoint
`EMBEDDING_MODEL`	`bge-m3`	Model for embeddings
`CHAT_MODEL`	`qwen3:0.6b`	Model for AI Q&A
`ENABLE_AI_CHAT`	`false`	Enable AI Q&A feature
`COMPACT_ENABLED`	`false`	Enable LLM Compact feature

FTS Tokenizer

Controls how full-text search indexes text. Configure via compact.fts_tokenizer in config file.

Tokenizer	Language	Matching	Index Size
`trigram` (default)	CJK + English	Substring	Larger
`unicode61`	English only	Word boundary	Smaller

Use trigram if you search in Chinese, Japanese, or Korean. Use unicode61 for English-only projects with smaller index size.

Compact (Experimental)

LLM-powered session summarization. Inspired by claude-mem. Disabled by default.

{
  "compact": {
    "enabled": true,
    "l2_talk_summary": true,
    "l3_session_summary": true
  }
}

Requires a chat model (Ollama or cloud provider). When enabled, generates summaries for MCP search results.

Models

Embedding: bge-m3 (default). For faster/smaller: nomic-embed-text
Chat: qwen3:0.6b (default). For faster/smaller: llama3.2:3b

Verify

# Health check
curl http://localhost:10013/health

# Stats
curl http://localhost:10013/api/stats

# Embedding status
curl http://localhost:10013/api/embedding/status

Troubleshooting

Semantic search not working

# Check Ollama
curl http://localhost:11434/api/tags

# Check model
ollama list

No sessions found

# Check data paths exist
ls ~/.claude/projects/
ls ~/.codex/
ls ~/.local/share/opencode/
ls ~/.gemini/tmp/

# Trigger manual collection
curl -X POST http://localhost:10013/api/collect