Struct AnthropicCreateMessageRequest 

pub struct AnthropicCreateMessageRequest {

    pub model: String,
    pub max_tokens: u32,
    pub messages: Vec<AnthropicMessage>,
    pub nvext: Option<Value>,
    pub system: Option<SystemContent>,
    pub temperature: Option<f32>,
    pub top_p: Option<f32>,
    pub top_k: Option<u32>,
    pub stop_sequences: Option<Vec<String>>,
    pub stream: bool,
    pub metadata: Option<Value>,
    pub tools: Option<Vec<AnthropicTool>>,
    pub tool_choice: Option<AnthropicToolChoice>,
    pub cache_control: Option<CacheControl>,
    pub thinking: Option<ThinkingConfig>,
    pub service_tier: Option<String>,
    pub container: Option<String>,
    pub output_config: Option<Value>,
}

Top-level request body for POST /v1/messages.

Fields

model: String

The model to use (e.g. “claude-sonnet-4-20250514”).

max_tokens: u32

The maximum number of tokens to generate.

messages: Vec<AnthropicMessage>

The conversation messages.

nvext: Option<Value>

Dynamo protocol extension envelope. Protocol parsing keeps this opaque; LLM request handling owns extension validation and normalization.

system: Option<SystemContent>

Optional system prompt (string or array of {"type":"text","text":"..."} blocks).

temperature: Option<f32>

Sampling temperature (0.0 - 1.0).

top_p: Option<f32>

Nucleus sampling parameter.

top_k: Option<u32>

Top-K sampling parameter.

stop_sequences: Option<Vec<String>>

Custom stop sequences.

stream: bool

Whether to stream the response.

metadata: Option<Value>

Optional metadata (e.g. user_id).

tools: Option<Vec<AnthropicTool>>

Tools the model may call.

tool_choice: Option<AnthropicToolChoice>

How the model should choose which tool to call.

cache_control: Option<CacheControl>

Top-level cache control for automatic prompt prefix caching. When present, the system caches all content up to the last cacheable block. Matches the Anthropic Messages API automatic caching mode. See: https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching#automatic-caching

thinking: Option<ThinkingConfig>

Extended thinking configuration. When enabled, the model produces thinking content blocks containing its internal reasoning before the final response. The budget_tokens field controls how many tokens the model may use for thinking (must be >= 1024 and < max_tokens).

service_tier: Option<String>

Service tier selection: "auto" or "standard_only".

container: Option<String>

Container identifier for stateful sandbox sessions.

output_config: Option<Value>

Output configuration: effort level and optional JSON schema format. effort can be "low", "medium", "high", or "max". format specifies structured JSON output constraints.

Implementations

impl AnthropicCreateMessageRequest

pub fn estimate_tokens(&self) -> u32

Estimate input token count using the same len/3 heuristic as /count_tokens.

Trait Implementations

impl Clone for AnthropicCreateMessageRequest

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for AnthropicCreateMessageRequest

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for AnthropicCreateMessageRequest

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for AnthropicCreateMessageRequest

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

where __S: Serializer,

Serialize this value into the given Serde serializer.