Enum NormalizedEvent 

pub enum NormalizedEvent {
    StreamStart {
        request_id: String,
    },
    MessageDelta {
        text: String,
    },
    ThinkingDelta {
        text: String,
    },
    ReasoningDelta {
        text: String,
    },
    CitationAdded(Citation),
    MemoryUpdate {
        key: String,
        value: String,
        operation: String,
    },
    SkillActivation {
        name: String,
        status: String,
    },
    Custom {
        source: String,
        event_name: String,
        payload: Value,
    },
    ToolCallDelta {
        call_index: usize,
        id: Option<String>,
        name: Option<String>,
        arguments_delta: Option<String>,
    },
    ToolCallComplete {
        call_index: usize,
        id: String,
        name: String,
        arguments_json: String,
    },
    ToolResult {
        id: String,
        name: String,
        content: String,
        success: bool,
    },
    SandboxOutput {
        stream: String,
        data: String,
    },
    Error {
        message: String,
        code: Option<String>,
    },
    Usage {
        prompt_tokens: u32,
        completion_tokens: u32,
        total_tokens: u32,
        cached_tokens: Option<u32>,
        cache_creation_tokens: Option<u32>,
    },
    RuntimeStep {
        step: u32,
        kind: RuntimeStepKind,
    },
    Done,
}

Normalized streaming events emitted by the LLM orchestrator.

These events provide a unified interface for the client UI regardless of which LLM protocol is used (Chat Completions, Responses API, etc.).

Variants

StreamStart

Indicates the start of a new streaming response.

Fields

request_id: String

Unique identifier for this request/response pair.

MessageDelta

Incremental text delta from the assistant’s response.

Fields

text: String

The text fragment to append.

ThinkingDelta

Incremental thinking/internal reasoning delta (for models that expose this).

Fields

text: String

The thinking text fragment to append.

ReasoningDelta

Incremental reasoning delta (chain-of-thought output).

Fields

text: String

The reasoning text fragment to append.

CitationAdded(Citation)

A citation/source reference was added.

MemoryUpdate

Memory/context update from the model.

Fields

key: String

Key for the memory entry.

value: String

Value to store.

operation: String

Operation type: “set”, “append”, or “delete”.

SkillActivation

A skill or skill-like provider capability was activated.

UAR proper owns real skill routing/execution. This lower-level event is for external/local drivers that need to preserve provider-emitted skill activation chunks without flattening them into assistant prose.

Fields

name: String

Skill identifier or display name.

status: String

Lifecycle/status label such as activated, running, complete, or failed.

Custom

Provider-specific semantic event that should remain inspectable.

Fields

source: String

Provider/source namespace.

event_name: String

Provider event name.

payload: Value

Original payload.

ToolCallDelta

Incremental tool call delta (streaming tool call assembly).

Fields

call_index: usize

Index of this tool call in the current batch.

id: Option<String>

Tool call ID (may arrive in first delta or later).

name: Option<String>

Tool/function name (may arrive in first delta or later).

arguments_delta: Option<String>

Incremental arguments JSON fragment.

ToolCallComplete

Tool call is fully assembled and ready for execution.

Fields

call_index: usize

Index of this tool call in the current batch.

id: String

Tool call ID.

name: String

Tool/function name.

arguments_json: String

Complete arguments as JSON string.

ToolResult

Result from executing a tool.

Fields

id: String

Tool call ID this result corresponds to.

name: String

Tool/function name.

content: String

Result content (typically JSON).

success: bool

Whether the tool execution succeeded.

SandboxOutput

Real-time output from a sandboxed code execution.

Fields

stream: String

Output stream: "stdout" or "stderr".

data: String

Output data fragment.

Error

An error occurred during streaming.

Fields

message: String

Error message.

code: Option<String>

Optional error code for programmatic handling.

Usage

Token usage information from the API.

Fields

prompt_tokens: u32

Total prompt/input tokens, including cache reads and cache writes.

completion_tokens: u32

Number of tokens in the completion/output.

total_tokens: u32

Total tokens used (prompt + completion).

cached_tokens: Option<u32>

Tokens served from the provider’s prompt cache (OpenAI: cached_tokens; Anthropic: cache_read_input_tokens).

cache_creation_tokens: Option<u32>

Prompt tokens written into the provider’s cache, already included in prompt_tokens (Anthropic: cache_creation_input_tokens).

RuntimeStep

A tool-loop iteration boundary — per-step run progress. Emitted at the start and end of each orchestrator iteration so the Runtime Console can show step progress.

Fields

step: u32

Monotonic per-run step index (the orchestrator iteration number).

kind: RuntimeStepKind

Whether the step is beginning or ending.

Done

Stream has completed successfully.

Trait Implementations

impl Clone for NormalizedEvent

fn clone(&self) -> NormalizedEvent

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for NormalizedEvent

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for NormalizedEvent

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for NormalizedEvent

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

Equality operator ==.

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

Inequality operator !=.

impl Serialize for NormalizedEvent

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for NormalizedEvent