Struct CostRecord 

pub struct CostRecord {

    pub id: String,
    pub channel_id: String,
    pub upstream_channel: String,
    pub upstream_model: String,
    pub request_time_ms: i64,
    pub project: String,
    pub user_id: String,
    pub agent_type: String,
    pub input_tokens: i64,
    pub output_tokens: i64,
    pub cache_write_tokens: i64,
    pub cache_read_tokens: i64,
    pub thinking_tokens: i64,
    pub cost: f64,
    pub schema_saved_tokens: i64,
    pub response_saved_tokens: i64,
    pub rtk_saved_tokens: i64,
    pub pre_compress_tokens: i64,
    pub post_compress_tokens: i64,
    pub compression_tokens_saved: i64,
    pub unit: String,
    pub pricing_snapshot_json: String,
    pub timestamp: String,
    pub session_id: Option<String>,
    pub before_tokens: i64,
    pub after_tokens: i64,
    pub tokens_saved: i64,
    pub compression_breakdown_json: String,
}

A single proxied request with token usage and cost.

Fields

id: String

UUID v7 primary key.

channel_id: String

Channel used for this request (proxy channel ID, e.g. “deepseek”).

upstream_channel: String

Human-readable upstream channel name (e.g. “DeepSeek Official”).

upstream_model: String

Upstream model name sent to the API (e.g. “deepseek-v4-pro”).

request_time_ms: i64

Request processing time in milliseconds (from request arrival to response return).

project: String

Project path or identifier.

user_id: String

User who made the request.

agent_type: String

Agent type: “ClaudeCode”, “Codex”, etc.

input_tokens: i64

Input/prompt tokens consumed.

output_tokens: i64

Output/completion tokens consumed.

cache_write_tokens: i64

Tokens written to the provider’s prompt cache.

cache_read_tokens: i64

Tokens read from the provider’s prompt cache.

thinking_tokens: i64

Extended thinking tokens consumed.

cost: f64

Actual monetary cost of this request.

schema_saved_tokens: i64

Tokens saved by schema compression.

response_saved_tokens: i64

Tokens saved by response compression.

rtk_saved_tokens: i64

Tokens saved by RTK token optimization.

pre_compress_tokens: i64

Token count before tokenless compression.

post_compress_tokens: i64

Token count after tokenless compression.

compression_tokens_saved: i64

Tokens saved by tokenless compression.

unit: String

Currency of cost: “USD”, “CNY”, “credits”.

pricing_snapshot_json: String

Serialized pricing snapshot for audit trail.

timestamp: String

When the request was completed (RFC 3339).

session_id: Option<String>

Session ID from X-Claude-Code-Session-Id header (for billing correlation).

before_tokens: i64

Estimated tokens before any compression (tokenless + proxy layers).

after_tokens: i64

Actual tokens consumed by the upstream API (input + output).

tokens_saved: i64

Total tokens saved across all compression layers.

compression_breakdown_json: String

JSON array breakdown of each compression operation.

Each element:

{
  "op": "compress-schema",      // opType — see report.rs for full enum
  "method": "ToonHrv",          // compression strategy
  "beforeTokens": 1500,
  "afterTokens": 700,
  "savedTokens": 800,
  "beforeBytes": 6000,
  "afterBytes": 2800,
  "savedBytes": 3200
}

See [crate::report] module docs for the complete opType / method lookup table.

Trait Implementations

impl Clone for CostRecord

fn clone(&self) -> CostRecord

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CostRecord

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for CostRecord

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for CostRecord

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

where __S: Serializer,

Serialize this value into the given Serde serializer.