Configuration

Graphmind is configured through three layers (later layers override earlier ones):

1. Config file (TOML) -- defaults to graphmind.toml in the working directory
2. CLI flags -- override config file values
3. Environment variables -- override everything

Config File

Create a graphmind.toml:

[server]
host = "127.0.0.1"       # Bind address
resp_port = 6379          # RESP protocol port
http_port = 8080          # HTTP API and Visualizer port
data_dir = "./graphmind_data"  # RocksDB data directory

[logging]
level = "info"            # debug, info, warn, error

[auth]
enabled = false           # Enable token-based authentication
tokens = [                # Allowed bearer tokens
  "my-secret-token-1",
  "my-secret-token-2",
]

Specify a config file path:

graphmind --config /etc/graphmind/graphmind.toml

If the file does not exist, Graphmind starts with defaults (no error).

CLI Flags

Flag	Config key	Default	Description
--config <path>	--	graphmind.toml	Config file path
--host <addr>	server.host	127.0.0.1	Bind address
--port <port>	server.resp_port	6379	RESP server port
--http-port <port>	server.http_port	8080	HTTP server port
--data-dir <path>	server.data_dir	./graphmind_data	Data directory
--log-level <level>	logging.level	info	Log level
--demo <mode>	--	(none)	Load demo data on startup

Environment Variables

Variable	Config key	Description
GRAPHMIND_HOST	server.host	Bind address
GRAPHMIND_PORT	server.resp_port	RESP port
GRAPHMIND_HTTP_PORT	server.http_port	HTTP port
GRAPHMIND_DATA_DIR	server.data_dir	Data directory
GRAPHMIND_AUTH_TOKEN	auth.tokens	Single auth token (enables auth)
RUST_LOG	logging.level	Log level filter

NLQ Provider Variables

These enable natural language query translation. Set one:

Variable	Provider	Default Model
OPENAI_API_KEY	OpenAI	gpt-4o-mini (override with OPENAI_MODEL)
GEMINI_API_KEY	Google Gemini	gemini-2.0-flash (override with GEMINI_MODEL)
CLAUDE_CODE_NLQ=1	Claude Code CLI	Uses local Claude Code

Example Configurations

Development

[server]
host = "127.0.0.1"
resp_port = 6379
http_port = 8080

[logging]
level = "debug"

Production (Single Node)

[server]
host = "0.0.0.0"
resp_port = 6379
http_port = 8080
data_dir = "/var/lib/graphmind/data"

[logging]
level = "info"

[auth]
enabled = true
tokens = ["prod-token-abc123"]

Docker / Kubernetes

When running in a container, use environment variables instead of a config file:

GRAPHMIND_HOST=0.0.0.0
GRAPHMIND_PORT=6379
GRAPHMIND_HTTP_PORT=8080
GRAPHMIND_DATA_DIR=/data
GRAPHMIND_AUTH_TOKEN=prod-token-abc123
RUST_LOG=info

Precedence

If the same setting is specified in multiple places, the order is:

Environment variable > CLI flag > Config file > Default

For example, if the config file says resp_port = 6379, the CLI flag --port 7379 overrides it, and GRAPHMIND_PORT=8379 overrides both.