Struct EngineConfig 

pub struct EngineConfig {
    pub kind: EngineKind,
    pub url: String,
    pub kv_offload: KvOffload,
    pub vllm: VllmEngineConfig,
    pub sglang: SglangEngineConfig,
    pub llama_cpp: LlamaCppEngineConfig,
}

Inference engine driver.

Cognitora’s cgn-agent proxies to an OpenAI-compatible HTTP server. This block describes which engine to spawn and how. Four kinds are supported:

• vllm — the agent spawns vllm serve <model> ... (GPU).
• sglang — the agent spawns python -m sglang.launch_server ... (GPU). SGLang offers RadixAttention prefix caching and structured-output acceleration; from the router’s perspective it speaks the same OpenAI surface as vLLM and is fully interchangeable.
• llama_cpp — the agent spawns python -m llama_cpp.server or a standalone llama-server binary (CPU or GPU offload).
• openai_compat — the agent does not spawn anything; it just proxies to engine.url. Use this when the engine is managed by systemd / Kubernetes / a sidecar.

Fields

kind: EngineKind

url: String

HTTP base URL where the engine exposes the OpenAI surface.

kv_offload: KvOffload

Engine-side KV offload / cross-worker connector. Picks the --kv-transfer-config JSON for vLLM and the --enable-hierarchical-cache flag set for SGLang. See KvOffload.

vllm: VllmEngineConfig

vLLM-specific knobs (used when kind = "vllm").

sglang: SglangEngineConfig

SGLang-specific knobs (used when kind = "sglang").

llama_cpp: LlamaCppEngineConfig

llama.cpp-specific knobs (used when kind = "llama_cpp").

Trait Implementations

impl Clone for EngineConfig

fn clone(&self) -> EngineConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for EngineConfig

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

Formats the value using the given formatter.

impl Default for EngineConfig

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for EngineConfig

where EngineConfig: Default,

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for EngineConfig

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

where __S: Serializer,

Serialize this value into the given Serde serializer.