Struct MessageCreateParams 

pub struct MessageCreateParams {

    pub max_tokens: u32,
    pub messages: Vec<MessageParam>,
    pub model: Model,
    pub metadata: Option<Metadata>,
    pub output_format: Option<OutputFormat>,
    pub stop_sequences: Option<Vec<String>>,
    pub system: Option<SystemPrompt>,
    pub temperature: Option<f32>,
    pub thinking: Option<ThinkingConfig>,
    pub tool_choice: Option<ToolChoice>,
    pub tools: Option<Vec<ToolUnionParam>>,
    pub top_k: Option<u32>,
    pub top_p: Option<f32>,
    pub stream: bool,
    pub effort: Option<EffortLevel>,
    pub speed: Option<SpeedMode>,
    pub output_config: Option<OutputConfig>,
    pub context_management: Option<ContextManagement>,
    pub inference_geo: Option<String>,
    pub service_tier: Option<String>,
    pub container: Option<String>,
    pub cache_control: Option<CacheControlEphemeral>,
    pub citations: Option<CitationsConfig>,
    pub skills: Option<Vec<SkillRef>>,
    pub tool_runner: Option<bool>,
}

Available on crate feature anthropic-client only.

Parameters for creating messages.

Fields

max_tokens: u32

The maximum number of tokens to generate before stopping.

Note that our models may stop before reaching this maximum. This parameter only specifies the absolute maximum number of tokens to generate.

Different models have different maximum values for this parameter. See models for details.

messages: Vec<MessageParam>

Input messages.

Our models are trained to operate on alternating user and assistant conversational turns. When creating a new Message, you specify the prior conversational turns with the messages parameter, and the model then generates the next Message in the conversation. Consecutive user or assistant turns in your request will be combined into a single turn.

model: Model

The model that will complete your prompt.

See models for additional details and options.

metadata: Option<Metadata>

An object describing metadata about the request.

output_format: Option<OutputFormat>

Output format configuration for structured outputs (legacy field).

Prefer output_config for new code. This field is retained for backward compatibility with existing callers.

stop_sequences: Option<Vec<String>>

Custom text sequences that will cause the model to stop generating.

system: Option<SystemPrompt>

System prompt.

temperature: Option<f32>

Amount of randomness injected into the response.

Defaults to 1.0. Ranges from 0.0 to 1.0.

thinking: Option<ThinkingConfig>

Configuration for enabling Claude’s extended thinking.

tool_choice: Option<ToolChoice>

How the model should use the provided tools.

tools: Option<Vec<ToolUnionParam>>

Definitions of tools that the model may use.

top_k: Option<u32>

Only sample from the top K options for each subsequent token.

top_p: Option<f32>

Use nucleus sampling.

stream: bool

Whether to incrementally stream the response using server-sent events.

effort: Option<EffortLevel>

Top-level effort parameter controlling response thoroughness.

speed: Option<SpeedMode>

Speed mode for latency-critical workloads (research preview).

output_config: Option<OutputConfig>

Structured output configuration (replaces output_format).

context_management: Option<ContextManagement>

Server-side context management (tool result clearing, thinking block clearing). Requires beta header context-management-2025-06-27.

inference_geo: Option<String>

Geographic routing for data residency control (e.g. “US”, “EU”).

service_tier: Option<String>

Service tier for priority capacity ("auto" or "standard_only").

container: Option<String>

Container identifier for code execution tool reuse across requests.

cache_control: Option<CacheControlEphemeral>

Top-level cache control. When set, the server automatically caches everything up to the last cacheable block.

citations: Option<CitationsConfig>

Citations configuration.

skills: Option<Vec<SkillRef>>

Skill references for dynamic instruction loading.

tool_runner: Option<bool>

Opt into programmatic tool calling flow.

Implementations

impl MessageCreateParams

pub fn new( max_tokens: u32, messages: Vec<MessageParam>, model: Model, ) -> MessageCreateParams

Create a new message creation parameters with streaming disabled.

pub fn new_streaming( max_tokens: u32, messages: Vec<MessageParam>, model: Model, ) -> MessageCreateParams

Create new streaming message creation parameters.

pub fn with_metadata(self, metadata: Metadata) -> MessageCreateParams

Add metadata to the parameters.

pub fn with_output_format( self, output_format: OutputFormat, ) -> MessageCreateParams

Add output format for structured outputs.

When set, constrains Claude’s response to follow a specific JSON schema, ensuring valid, parseable output for downstream processing.

This feature requires the beta header structured-outputs-2025-11-13.

Example

use serde_json::json;
use adk_anthropic::{MessageCreateParams, OutputFormat, KnownModel};

let params = MessageCreateParams::simple("Extract info", KnownModel::ClaudeHaiku45)
    .with_output_format(OutputFormat::json_schema(json!({
        "type": "object",
        "properties": {
            "name": { "type": "string" }
        },
        "required": ["name"],
        "additionalProperties": false
    })));

pub fn with_stop_sequences( self, stop_sequences: Vec<String>, ) -> MessageCreateParams

Add stop sequences to the parameters.

pub fn with_system_string(self, system: String) -> MessageCreateParams

Add a system prompt as a string.

pub fn with_system_blocks(self, blocks: Vec<TextBlock>) -> MessageCreateParams

Add a system prompt as text blocks.

pub fn with_system(self, system: impl Into<SystemPrompt>) -> MessageCreateParams

Add a system prompt.

pub fn with_temperature( self, temperature: f32, ) -> Result<MessageCreateParams, Error>

Add temperature to the parameters.

pub fn with_thinking(self, thinking: ThinkingConfig) -> MessageCreateParams

Add thinking configuration to the parameters.

pub fn with_tool_choice(self, tool_choice: ToolChoice) -> MessageCreateParams

Add tool choice to the parameters.

pub fn with_tools(self, tools: Vec<ToolUnionParam>) -> MessageCreateParams

Add tools to the parameters.

pub fn with_top_k(self, top_k: u32) -> MessageCreateParams

Add top_k to the parameters.

pub fn with_top_p(self, top_p: f32) -> Result<MessageCreateParams, Error>

Add top_p to the parameters.

pub fn with_stream(self, stream: bool) -> MessageCreateParams

Sets the streaming option.

pub fn validate(&self) -> Result<(), Error>

Validate all parameters before sending to the API with security checks.

Performs comprehensive validation including DoS prevention measures:

• Limits on message count and size to prevent resource exhaustion
• Validation of all numeric ranges
• Security checks on string inputs

pub fn simple( prompt: impl Into<MessageParam>, model: impl Into<Model>, ) -> MessageCreateParams

Create a simple message request with sensible defaults.

This is a convenience method for creating basic message requests without needing to specify all parameters explicitly.

pub fn simple_streaming( prompt: impl Into<MessageParam>, model: impl Into<Model>, ) -> MessageCreateParams

Create a simple streaming message request with sensible defaults.

pub fn with_message( self, message: impl Into<MessageParam>, ) -> MessageCreateParams

Add a single message to the parameters.

pub fn with_messages( self, messages: impl IntoIterator<Item = impl Into<MessageParam>>, ) -> MessageCreateParams

Add multiple messages to the parameters.

pub fn requires_structured_outputs_beta(&self) -> bool

Check if this request requires the structured outputs beta header.

Returns true if either:

• output_format is set (for JSON outputs)
• Any tool has strict: true (for strict tool use)

When this returns true, the client should include the anthropic-beta: structured-outputs-2025-11-13 header.

Trait Implementations

impl Clone for MessageCreateParams

fn clone(&self) -> MessageCreateParams

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MessageCreateParams

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

Formats the value using the given formatter.

impl Default for MessageCreateParams

fn default() -> MessageCreateParams

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for MessageCreateParams

fn deserialize<__D>( __deserializer: __D, ) -> Result<MessageCreateParams, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for MessageCreateParams

fn eq(&self, other: &MessageCreateParams) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for MessageCreateParams

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for MessageCreateParams