Struct Config 

pub struct Config {

    pub last_used_model: Option<String>,
    pub default_model: ModelSettings,
    pub ollama: OllamaConfig,
    pub non_interactive: NonInteractiveConfig,
    pub mcp_servers: HashMap<String, McpServerConfig>,
    pub providers: HashMap<String, UserProviderConfig>,
    pub reasoning_per_model: HashMap<String, ReasoningLevel>,
    pub ollama_num_ctx_per_model: HashMap<String, u32>,
    pub model_profiles: HashMap<String, String>,
    pub safety: SafetyConfig,
    pub memory: MemoryConfig,
    pub compaction: CompactionConfig,
    pub computer_use: ComputerUseConfig,
    pub prompt: PromptConfig,
}

Main configuration structure

Fields

last_used_model: Option<String>

Last used model (persisted between sessions)

default_model: ModelSettings

Default model configuration

ollama: OllamaConfig

Ollama configuration

non_interactive: NonInteractiveConfig

Non-interactive mode configuration

mcp_servers: HashMap<String, McpServerConfig>

MCP server configurations

providers: HashMap<String, UserProviderConfig>

User overrides + custom OpenAI-compatible providers. Keys are provider names; matching a built-in registry entry overrides its defaults, anything else defines a fully custom provider. Example:

[providers.groq]
api_key_env = "MY_GROQ_KEY"  # override default GROQ_API_KEY

[providers.my-vllm]
base_url = "http://192.168.1.42:8000/v1"
api_key_env = "VLLM_KEY"
compat = "openai-effort"

reasoning_per_model: HashMap<String, ReasoningLevel>

Per-model reasoning preferences keyed by full model ID (provider/name). Set when the user runs /reasoning <level> or Alt+T cycles while using a specific model — the new value sticks for that model until changed. Falls back to default_model.reasoning when no entry exists. Example:

[reasoning_per_model]
"anthropic/claude-sonnet-4-6" = "high"
"ollama/qwen3-coder:30b" = "low"

ollama_num_ctx_per_model: HashMap<String, u32>

Per-model Ollama num_ctx override set via /context <n>/max. Beats auto-fit; cleared by /context auto. Keyed by model id.

Example:

[ollama_num_ctx_per_model]
"ollama/ornith:9b" = 131072

model_profiles: HashMap<String, String>

Named model profiles that agents/plugins can request without hardcoding a concrete provider model. Values are full model IDs. Example:

[model_profiles]
fast = "ollama/qwen3-coder:14b"
large-context = "openai/gpt-5.2"
tool-strong = "anthropic/claude-sonnet-4-6"
vision = "gemini/gemini-2.5-pro"
cheap = "groq/llama-3.3-70b-versatile"

safety: SafetyConfig

Runtime safety policy. Defaults to Ask so mutations / shell / network actions require approval out of the box; users opt into Auto (LLM-vetted) or FullAccess deliberately.

memory: MemoryConfig

Durable semantic memory settings.

compaction: CompactionConfig

Context-compaction settings.

computer_use: ComputerUseConfig

Computer-use (desktop control) preferences.

prompt: PromptConfig

Runtime-only prompt customizations supplied by CLI flags. These are deliberately skipped when saving config so one-off agent personas do not pollute the user’s persistent Mermaid settings.

Trait Implementations

impl Clone for Config

fn clone(&self) -> Config

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Config

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for Config

fn default() -> Config

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for Config

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for Config

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.