Struct OpenAiCompat 

#[non_exhaustive]
pub struct OpenAiCompat {
    pub supports_store: bool,
    pub supports_developer_role: bool,
    pub supports_reasoning_effort: bool,
    pub supports_thinking_control: bool,
    pub supports_usage_in_streaming: bool,
    pub max_tokens_field: MaxTokensField,
    pub requires_tool_result_name: bool,
    pub requires_assistant_after_tool_result: bool,
    pub thinking_format: ThinkingFormat,
    pub supports_prompt_cache_key: bool,
}

Compatibility flags for OpenAI-compatible providers. Different providers have different quirks even though they share the same base API.

Marked #[non_exhaustive]: this is the crate’s most literal instance of a growing quirk list — every new provider difference adds a flag, and without the attribute each one is a downstream break. Construct from a preset (OpenAiCompat::openai, OpenAiCompat::deepseek, …) or Default::default and adjust fields. New flags carry #[serde(default)] so persisted configs keep deserializing.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

supports_store: bool

Supports the store parameter for conversation persistence.

supports_developer_role: bool

Supports developer role (system-level instructions).

supports_reasoning_effort: bool

Supports reasoning_effort parameter.

supports_thinking_control: bool

Supports DeepSeek-style thinking mode control.

supports_usage_in_streaming: bool

Includes usage data in streaming responses.

max_tokens_field: MaxTokensField

Which field name to use for max tokens.

requires_tool_result_name: bool

Tool results must include a name field.

requires_assistant_after_tool_result: bool

Must insert an assistant message after tool results.

thinking_format: ThinkingFormat

How thinking/reasoning content is formatted in streaming.

supports_prompt_cache_key: bool

Accepts OpenAI’s prompt_cache_key for routing cache lookups.

Off by default: the field is OpenAI’s, and a strict compat server that validates unknown keys would reject the whole request rather than ignore it. Providers that cache automatically (DeepSeek, Groq) lose nothing by leaving this off — they were never reading it.

Implementations

impl OpenAiCompat

pub fn openai() -> Self

Compat flags for native OpenAI.

pub fn meta() -> Self

Compat flags for the Meta Model API (Muse Spark).

OpenAI-compatible chat completions. Meta documents reasoning_effort (default medium server-side) and streamed usage via stream_options.include_usage; max_tokens is deprecated in favor of max_completion_tokens.

pub fn xai() -> Self

Compat flags for xAI (Grok).

pub fn groq() -> Self

Compat flags for Groq.

pub fn cerebras() -> Self

Compat flags for Cerebras.

pub fn openrouter() -> Self

Compat flags for OpenRouter.

pub fn mistral() -> Self

Compat flags for Mistral.

pub fn deepseek() -> Self

Compat flags for DeepSeek.

pub fn zai() -> Self

Compat flags for Z.ai (Zhipu AI).

pub fn minimax() -> Self

Compat flags for MiniMax.

pub fn qwen() -> Self

Compat flags for Qwen / DashScope.

pub fn ollama() -> Self

Compat flags for Ollama’s OpenAI-compatible API.

Trait Implementations

impl Clone for OpenAiCompat

fn clone(&self) -> OpenAiCompat

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for OpenAiCompat

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

Formats the value using the given formatter.

impl Default for OpenAiCompat

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for OpenAiCompat

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for OpenAiCompat

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

where __S: Serializer,

Serialize this value into the given Serde serializer.