### Install RelayKit via Go CLI

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit

Use the following command to install the RelayKit module in your Go project.

```bash
go get github.com/QuantumNous/new-api/relaykit@latest
```

--------------------------------

### Convert OpenAI Chat Request to Claude Messages

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit

This example demonstrates how to initialize a request DTO and use ConvertRequest to transform an OpenAI Chat Completions request into a Claude Messages format.

```go
package main

import (
	"context"
	"fmt"

	"github.com/QuantumNous/new-api/relaykit/dto"
	"github.com/QuantumNous/new-api/relaykit/relayconvert"
	"github.com/QuantumNous/new-api/relaykit/relayconvert/convmeta"
	"github.com/QuantumNous/new-api/relaykit/types"
)

func main() {
	maxTokens := uint(1024)
	request := &dto.GeneralOpenAIRequest{
		Model: "claude-sonnet-4-5",
		Messages: []dto.Message{
			{Role: "user", Content: "Hello!"},
		},
		MaxTokens: &maxTokens,
	}

	meta := &convmeta.Values{
		OriginModelName:     "client-model",
		UpstreamModelName:   request.Model,
		ChannelMetaAttached: true,
	}

	result, err := relayconvert.ConvertRequest(
		context.Background(),
		meta,
		types.RelayFormatClaude,
		request,
	)
	if err != nil {
		panic(err)
	}

	claudeRequest, ok := result.Value.(*dto.ClaudeRequest)
	if !ok {
		panic(fmt.Sprintf("unexpected result type %T", result.Value))
	}

	fmt.Printf("model=%s messages=%d\n", claudeRequest.Model, len(claudeRequest.Messages))
}
```

--------------------------------

### EquivalentGeminiStrength function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Compares two controls after applying the target model's budget/level mapping. Accepts distinct canonical labels that are identical on the Gemini wire (for example minimal and low on 2.5).

```go
func EquivalentGeminiStrength(model string, left Intent, right Intent) (bool, error)
```

--------------------------------

### ParseOpenAIReasoningEffortFromModelSuffix function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Extracts an OpenAI effort tail such as -high or -none. preserveEffortTail, when non-nil, keeps real model IDs whose names already end in those tokens (for example qwen-max).

```go
func ParseOpenAIReasoningEffortFromModelSuffix(modelName string, preserveEffortTail func(string) bool) (string, string)
```

--------------------------------

### ClaudeRequest GetStringSystem

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Retrieves the system prompt as a string.

```go
func (c *ClaudeRequest) GetStringSystem() string
```

--------------------------------

### OptionsOf Function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/convmeta

Returns m's conversion options, or empty defaults when m is nil.

```go
func OptionsOf(m Meta) *Options
```

--------------------------------

### WebSearchOptions struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Configuration options for performing web searches.

```go
type WebSearchOptions struct {
	SearchContextSize string          `json:"search_context_size,omitempty"`
	UserLocation      json.RawMessage `json:"user_location,omitempty"`
}
```

--------------------------------

### Run tests and build in the relaykit directory

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit

Execute these commands within the relaykit directory to ensure the module remains independently buildable.

```bash
GOWORK=off go test ./...
GOWORK=off go build ./...
```

--------------------------------

### ClaudeRequest SetStringSystem

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Sets the system prompt as a string.

```go
func (c *ClaudeRequest) SetStringSystem(system string)
```

--------------------------------

### Source type

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Defines the Source type as a string.

```go
type Source string
```

--------------------------------

### Set media resolver for cross-protocol image conversion

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit

Configure media resolution callbacks at startup to handle URL content or data URLs. RelayKit will return an error if media resolution is required but these callbacks are not configured.

```go
relayconvert.SetMediaResolver(relayconvert.MediaResolver{
	GetBase64Data:        getBase64Data,
	DecodeBase64FileData: decodeBase64FileData,
})
```

--------------------------------

### Options Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/convmeta

The per-request snapshot of host configuration that converters consult. Zero value means every adaptation is disabled.

```go
type Options struct {
	Claude ClaudeOptions
	Gemini GeminiOptions

	// ToolLossPolicy controls whether a cross-protocol conversion may omit or
	// approximate built-in-tool semantics. The zero value uses the allow
	// policy: conversion succeeds and every loss is returned as a diagnostic.
	// safe/strict rejection is request-phase opt-in only; response and stream
	// conversion never reject regardless of this field.
	ToolLossPolicy types.ConversionLossPolicy

	// OpenRouterDialect marks the upstream as OpenRouter's OpenAI-compatible
	// surface, which accepts extra fields (reasoning config, cache_control on
	// system parts) that converters emit only for that dialect. The host sets
	// it from the channel type.
	OpenRouterDialect bool

	// PreserveThinkingSuffix reports models whose -thinking/-nothinking/effort
	// suffix must be kept on the outgoing model name (host blacklist lookup).
	// Nil means "never preserve".
	PreserveThinkingSuffix func(modelName string) bool

	// PreserveEffortTail reports real model IDs whose names already end in an
	// effort-like token (for example qwen-max). Nil means "never preserve".
	PreserveEffortTail func(modelName string) bool
}
```

--------------------------------

### UserSetting struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines user-specific configuration settings including notification preferences, webhook details, and UI preferences.

```go
type UserSetting struct {
	NotifyType                       string  `json:"notify_type,omitempty"`                          // QuotaWarningType 额度预警类型
	QuotaWarningThreshold            float64 `json:"quota_warning_threshold,omitempty"`              // QuotaWarningThreshold 额度预警阈值
	WebhookUrl                       string  `json:"webhook_url,omitempty"`                          // WebhookUrl webhook地址
	WebhookSecret                    string  `json:"webhook_secret,omitempty"`                       // WebhookSecret webhook密钥
	NotificationEmail                string  `json:"notification_email,omitempty"`                   // NotificationEmail 通知邮箱地址
	BarkUrl                          string  `json:"bark_url,omitempty"`                             // BarkUrl Bark推送URL
	GotifyUrl                        string  `json:"gotify_url,omitempty"`                           // GotifyUrl Gotify服务器地址
	GotifyToken                      string  `json:"gotify_token,omitempty"`                         // GotifyToken Gotify应用令牌
	GotifyPriority                   int     `json:"gotify_priority"`                                // GotifyPriority Gotify消息优先级
	UpstreamModelUpdateNotifyEnabled bool    `json:"upstream_model_update_notify_enabled,omitempty"` // 是否接收上游模型更新定时检测通知（仅管理员）
	AcceptUnsetRatioModel            bool    `json:"accept_unset_model_ratio_model,omitempty"`       // AcceptUnsetRatioModel 是否接受未设置价格的模型
	RecordIpLog                      bool    `json:"record_ip_log,omitempty"`                        // 是否记录请求和错误日志IP
	SidebarModules                   string  `json:"sidebar_modules,omitempty"`                      // SidebarModules 左侧边栏模块配置
	BillingPreference                string  `json:"billing_preference,omitempty"`                   // BillingPreference 扣费策略（订阅/钱包）
	Language                         string  `json:"language,omitempty"`                             // Language 用户语言偏好 (zh, en)
}
```

--------------------------------

### Configure conversion context with convmeta.Values

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit

Use convmeta.Values to define options for specific models like Claude and Gemini. Ensure Claude.DefaultMaxTokens is set when converting to Claude to avoid errors.

```go
meta := &convmeta.Values{
	Options: &convmeta.Options{
		Claude: convmeta.ClaudeOptions{
			DefaultMaxTokens: func(model string) int {
				return 4096
			},
		},
		Gemini: convmeta.GeminiOptions{
			ThinkingAdapterEnabled: true,
		},
	},
}
```

--------------------------------

### GetBudgetTokens method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Retrieves the budget tokens for the Thinking configuration.

```go
func (c *Thinking) GetBudgetTokens() int
```

--------------------------------

### HTTP Protocol Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants for HTTP protocol configuration and connection management.

```go
const (
	HTTPProtocolAuto         = "auto"
	HTTPProtocolHTTP1        = "http1"
	MaxHTTP2ConnectionShards = 8
)
```

--------------------------------

### Thinking struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Configuration for thinking process parameters, including budget tokens and display settings.

```go
type Thinking struct {
	Type         string `json:"type,omitempty"`
	BudgetTokens *int   `json:"budget_tokens,omitempty"`
	// Display controls whether thinking content is returned in the response.
	// Used with adaptive thinking on Claude Opus 4.7+: "summarized" restores
	// the visible summary that was default on Opus 4.6; "omitted" (default on
	// 4.7) suppresses it. Pass-through field from upstream Anthropic API.
	Display string `json:"display,omitempty"`
}
```

--------------------------------

### ConvOptions Method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/convmeta

Returns the request-scoped conversion options snapshot. Must never return nil.

```go
// ConvOptions returns the request-scoped conversion options snapshot.
	// Must never return nil.
	ConvOptions() *Options
}
```

--------------------------------

### EffortFromBudget function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Maps a numeric budget to an Effort level.

```go
func EffortFromBudget(budget int) Effort
```

--------------------------------

### GetStringContent method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Returns the content as a string. Use when the content field is a string or needs conversion.

```go
func (c *ClaudeMediaMessage) GetStringContent() string
```

--------------------------------

### GetContentString method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Returns the content as a string. Use when the content field is a string or needs conversion.

```go
func (c *ChatCompletionsStreamResponseChoiceDelta) GetContentString() string
```

--------------------------------

### Reasoning struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents reasoning metadata including effort, summary, and raw context.

```go
type Reasoning struct {
	Effort  string          `json:"effort,omitempty"`
	Summary string          `json:"summary,omitempty"`
	Mode    json.RawMessage `json:"mode,omitempty"`
	Context json.RawMessage `json:"context,omitempty"`
}
```

--------------------------------

### ClaudeRequest ParseSystem

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Parses the system prompt into a slice of ClaudeMediaMessage.

```go
func (c *ClaudeRequest) ParseSystem() []ClaudeMediaMessage
```

--------------------------------

### ToolConfig struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Configuration settings for tool-related features like function calling and retrieval.

```go
type ToolConfig struct {
	FunctionCallingConfig            *FunctionCallingConfig `json:"functionCallingConfig,omitempty"`
	RetrievalConfig                  *RetrievalConfig       `json:"retrievalConfig,omitempty"`
	IncludeServerSideToolInvocations *bool                  `json:"includeServerSideToolInvocations,omitempty"`
}
```

--------------------------------

### RenderClaude function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Renders a Claude reasoning configuration from an intent, applying the adapter budget percentage and max tokens.

```go
func RenderClaude(model string, intent Intent, maxTokens *uint, adapterBudgetPercentage float64) (ClaudeRender, error)
```

--------------------------------

### OutputTokenDetails Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Details regarding token usage across different modalities.

```go
type OutputTokenDetails struct {
	TextTokens      int `json:"text_tokens"`
	AudioTokens     int `json:"audio_tokens"`
	ImageTokens     int `json:"image_tokens"`
	ReasoningTokens int `json:"reasoning_tokens"`
}
```

--------------------------------

### Usage struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Detailed token usage statistics for API requests.

```go
type Usage struct {
	PromptTokens         int           `json:"prompt_tokens"`
	CompletionTokens     int           `json:"completion_tokens"`
	TotalTokens          int           `json:"total_tokens"`
	PromptCacheHitTokens int           `json:"prompt_cache_hit_tokens,omitempty"`
	UsageSemantic        string        `json:"usage_semantic,omitempty"`
	UsageSource          string        `json:"usage_source,omitempty"`
	BillingUsage         *BillingUsage `json:"billing_usage,omitempty"`

	PromptTokensDetails    InputTokenDetails  `json:"prompt_tokens_details"`
	CompletionTokenDetails OutputTokenDetails `json:"completion_tokens_details"`
	InputTokens            int                `json:"input_tokens"`
	OutputTokens           int                `json:"output_tokens"`
	InputTokensDetails     *InputTokenDetails `json:"input_tokens_details"`

	// claude cache 1h
	ClaudeCacheCreation5mTokens int `json:"claude_cache_creation_5_m_tokens"`
	ClaudeCacheCreation1hTokens int `json:"claude_cache_creation_1_h_tokens"`
```

--------------------------------

### OpenAIVideo Struct and Methods

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the video object structure and associated helper methods for metadata and progress management.

```go
type OpenAIVideo struct {
	ID                 string            `json:"id"`
	TaskID             string            `json:"task_id,omitempty"` //兼容旧接口 待废弃
	Object             string            `json:"object"`
	Model              string            `json:"model"`
	Status             string            `json:"status"` // Should use VideoStatus constants: VideoStatusQueued, VideoStatusInProgress, VideoStatusCompleted, VideoStatusFailed
	Progress           int               `json:"progress"`
	CreatedAt          int64             `json:"created_at"`
	CompletedAt        int64             `json:"completed_at,omitempty"`
	ExpiresAt          int64             `json:"expires_at,omitempty"`
	Seconds            string            `json:"seconds,omitempty"`
	Size               string            `json:"size,omitempty"`
	RemixedFromVideoID string            `json:"remixed_from_video_id,omitempty"`
	Error              *OpenAIVideoError `json:"error,omitempty"`
	Metadata           map[string]any    `json:"metadata,omitempty"`
}

func NewOpenAIVideo() *OpenAIVideo

func (m *OpenAIVideo) SetMetadata(k string, v any)

func (m *OpenAIVideo) SetProgressStr(progress string)
```

--------------------------------

### Built-in Tool Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants for built-in tools available in the system.

```go
const (
	BuildInToolWebSearchPreview = "web_search_preview"
	BuildInToolWebSearch        = "web_search"
	BuildInToolFileSearch       = "file_search"
	BuildInToolGoogleSearch     = "google_search"
	BuildInToolImageGeneration  = "image_generation"
)
```

--------------------------------

### MessageVideoUrl struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents a video URL input.

```go
type MessageVideoUrl struct {
	Url string `json:"url"`
}
```

--------------------------------

### GetReasoningContent method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Returns the reasoning content as a string. Use when reasoning content is present.

```go
func (c *ChatCompletionsStreamResponseChoiceDelta) GetReasoningContent() string
```

--------------------------------

### Tool struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines a tool available for invocation, including its schema and strictness.

```go
type Tool struct {
	Name        string                 `json:"name"`
	Description string                 `json:"description,omitempty"`
	InputSchema map[string]interface{} `json:"input_schema"`
	Strict      *bool                  `json:"strict,omitempty"`
}
```

--------------------------------

### OutputConfigForEffort Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Used for extracting effort configuration.

```go
type OutputConfigForEffort struct {
	Effort string `json:"effort,omitempty"`
}
```

--------------------------------

### IsQwenThinkingBudgetModel Function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Checks if the provided model name supports Qwen thinking budget (added in v0.2.0).

```go
func IsQwenThinkingBudgetModel(modelName string) bool
```

--------------------------------

### ClaudeCacheCreationUsage struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Tracks cache creation usage with ephemeral input token counts for 5-minute and 1-hour cache durations.

```go
type ClaudeCacheCreationUsage struct {
	Ephemeral5mInputTokens int `json:"ephemeral_5m_input_tokens,omitempty"`
	Ephemeral1hInputTokens int `json:"ephemeral_1h_input_tokens,omitempty"`
}
```

--------------------------------

### Billing Usage Source Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants defining the sources for billing usage tracking.

```go
const (
	BillingUsageSourceClaudeMessages = "claude_messages"
	BillingUsageSourceGeminiChat     = "gemini_chat"
	BillingUsageSourceOAIChat        = "oai_chat"
	BillingUsageSourceOAIResponses   = "oai_responses"

	BillingUsageSemanticAnthropic = "anthropic"
	BillingUsageSemanticGemini    = "gemini"
	BillingUsageSemanticOpenAI    = "openai"
)
```

--------------------------------

### RealTimeTool Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines a tool available in real-time interactions.

```go
type RealTimeTool struct {
	Type        string `json:"type"`
	Name        string `json:"name"`
	Description string `json:"description"`
	Parameters  any    `json:"parameters"`
}
```

--------------------------------

### RealtimeContent Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents content within a real-time message, supporting text and audio.

```go
type RealtimeContent struct {
	Type       string `json:"type"`
	Text       string `json:"text,omitempty"`
	Audio      string `json:"audio,omitempty"` // Base64-encoded audio bytes.
	Transcript string `json:"transcript,omitempty"`
}
```

--------------------------------

### GeminiThinkingConfig struct and methods

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Configuration for Gemini thinking capabilities, including methods for setting the budget and custom unmarshaling.

```go
type GeminiThinkingConfig struct {
	IncludeThoughts *bool `json:"includeThoughts,omitempty"`
	ThinkingBudget  *int  `json:"thinkingBudget,omitempty"`
	// TODO Conflict with thinkingbudget.
	ThinkingLevel string `json:"thinkingLevel,omitempty"`
}
```

```go
func (c *GeminiThinkingConfig) SetThinkingBudget(budget int)
```

```go
func (c *GeminiThinkingConfig) UnmarshalJSON(data []byte) error
```

--------------------------------

### ReasoningConversionState struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Carries provider-native reasoning controls between conversion steps. It is not part of the wire protocol and should be tagged with json:"-" in requests.

```go
type ReasoningConversionState struct {
	Mode            string
	Effort          string
	BudgetTokens    *int
	IncludeThoughts *bool
}
```

--------------------------------

### GeneralOpenAIRequest ParseInput method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Method signature for parsing input from a GeneralOpenAIRequest into a slice of strings.

```go
func (r *GeneralOpenAIRequest) ParseInput() []string
```

--------------------------------

### GetToolsMap Method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Returns the tools configuration as a slice of maps.

```go
func (r *OpenAIResponsesRequest) GetToolsMap() []map[string]any
```

--------------------------------

### ToFileSource method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Converts the message to a file source. Use to transform the message into a file source representation.

```go
func (c *ClaudeMediaMessage) ToFileSource()
```

--------------------------------

### Realtime Event Type Constants (Input)

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants for incoming realtime event types.

```go
const (
	RealtimeEventTypeError              = "error"
	RealtimeEventTypeSessionUpdate      = "session.update"
	RealtimeEventTypeConversationCreate = "conversation.item.create"
	RealtimeEventTypeResponseCreate     = "response.create"
	RealtimeEventInputAudioBufferAppend = "input_audio_buffer.append"
)
```

--------------------------------

### RealtimeSession Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents a session configuration in the real-time API.

```go
type RealtimeSession struct {
```

--------------------------------

### ClientError type

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Marks invalid user-supplied reasoning controls so host handlers can return a 4xx without classifying unrelated adapter failures as client errors.

```go
type ClientError struct {
	// contains filtered or unexported fields
}
```

--------------------------------

### GeminiChatTool type

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

GeminiChatTool struct for tool configuration, supporting GoogleSearch and GoogleSearchRetrieval.

```go
type GeminiChatTool struct {
	GoogleSearch          any             `json:"googleSearch,omitempty"`
	GoogleSearchRetrieval any             `json:"googleSearchRetrieval,omitempty"`
}
```

--------------------------------

### GeminiImage Structs

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Data structures for Imagen-related requests and responses.

```go
type GeminiImageInstance struct {
	Prompt string `json:"prompt"`
}
```

```go
type GeminiImageParameters struct {
	SampleCount      int    `json:"sampleCount,omitempty"`
	AspectRatio      string `json:"aspectRatio,omitempty"`
	PersonGeneration string `json:"personGeneration,omitempty"`
	ImageSize        string `json:"imageSize,omitempty"`
}
```

```go
type GeminiImagePrediction struct {
	MimeType           string `json:"mimeType"`
	BytesBase64Encoded string `json:"bytesBase64Encoded"`
	RaiFilteredReason  string `json:"raiFilteredReason,omitempty"`
	SafetyAttributes   any    `json:"safetyAttributes,omitempty"`
}
```

```go
type GeminiImageRequest struct {
	Instances  []GeminiImageInstance `json:"instances"`
	Parameters GeminiImageParameters `json:"parameters"`
}
```

```go
type GeminiImageResponse struct {
	Predictions []GeminiImagePrediction `json:"predictions"`
}
```

--------------------------------

### ChannelOtherSettings struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the ChannelOtherSettings struct with fields for Azure, Vertex, OpenRouter, and Claude-specific settings.

```go
type ChannelOtherSettings struct {
	AzureResponsesVersion                 string                `json:"azure_responses_version,omitempty"`
	VertexKeyType                         VertexKeyType         `json:"vertex_key_type,omitempty"` // "json" or "api_key"
	OpenRouterEnterprise                  *bool                 `json:"openrouter_enterprise,omitempty"`
	ClaudeBetaQuery                       bool                  `json:"claude_beta_query,omitempty"`          // Claude 渠道是否强制追加 ?beta=true
	AllowServiceTier                      bool                  `json:"allow_service_tier,omitempty"`         // 是否允许 service_tier 透传（默认过滤以避免额外计费）
	AllowInferenceGeo                     bool                  `json:"allow_inference_geo,omitempty"`        // 是否允许 inference_geo 透传（仅 Claude，默认过滤以满足数据驻留合规
	AllowSpeed                            bool                  `json:"allow_speed,omitempty"`                // 是否允许 speed 透传（仅 Claude，默认过滤以避免意外切换推理速度模式）
}
```

--------------------------------

### ClaudeRequest GetTools

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Retrieves the list of tools configured in the request.

```go
func (c *ClaudeRequest) GetTools() []any
```

--------------------------------

### NewGeminiChatBillingUsage function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Creates a BillingUsage from GeminiUsageMetadata.

```go
func NewGeminiChatBillingUsage(metadata *GeminiUsageMetadata) *BillingUsage
```

--------------------------------

### OpenAITextResponse Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents the standard text response structure from OpenAI.

```go
type OpenAITextResponse struct {
	Id      string                     `json:"id"`
	Model   string                     `json:"model"`
	Object  string                     `json:"object"`
	Created any                        `json:"created"`
	Choices []OpenAITextResponseChoice `json:"choices"`
	Error   any                        `json:"error,omitempty"`
	Usage   `json:"usage"`
}
```

--------------------------------

### ResponsesOutput struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents the output structure for responses, including content, reasoning, and tool-related fields.

```go
type ResponsesOutput struct {
	Type                string                          `json:"type"`
	ID                  string                          `json:"id"`
	Status              string                          `json:"status"`
	Role                string                          `json:"role"`
	Content             []ResponsesOutputContent        `json:"content"`
	Summary             []ResponsesReasoningSummaryPart `json:"summary,omitempty"`
	Quality             string                          `json:"quality"`
	Size                string                          `json:"size"`
	Result              string                          `json:"result,omitempty"`
	CallId              string                          `json:"call_id,omitempty"`
	Name                string                          `json:"name,omitempty"`
	Arguments           json.RawMessage                 `json:"arguments,omitempty"`
	Action              json.RawMessage                 `json:"action,omitempty"`
	Queries             json.RawMessage                 `json:"queries,omitempty"`
	Results             json.RawMessage                 `json:"results,omitempty"`
	Sources             json.RawMessage                 `json:"sources,omitempty"`
	Code                json.RawMessage                 `json:"code,omitempty"`
```

--------------------------------

### NewOpenAIResponsesBillingUsage function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Creates a BillingUsage from a Usage for OpenAI responses.

```go
func NewOpenAIResponsesBillingUsage(usage *Usage) *BillingUsage
```

--------------------------------

### Notification Type Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants for different types of system notifications.

```go
const (
	NotifyTypeQuotaExceed   = "quota_exceed"
	NotifyTypeChannelUpdate = "channel_update"
	NotifyTypeChannelTest   = "channel_test"
)
```

--------------------------------

### Intent.HasStrength method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Reports whether the intent has a strength setting.

```go
func (i Intent) HasStrength() bool
```

--------------------------------

### TextResponse struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Standard structure for text-based API responses.

```go
type TextResponse struct {
	Id      string                     `json:"id"`
	Object  string                     `json:"object"`
	Created int64                      `json:"created"`
	Model   string                     `json:"model"`
	Choices []OpenAITextResponseChoice `json:"choices"`
	Usage   `json:"usage"`
}
```

--------------------------------

### Video Status Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants representing the lifecycle status of video processing.

```go
const (
	VideoStatusUnknown    = "unknown"
	VideoStatusQueued     = "queued"
	VideoStatusInProgress = "in_progress"
	VideoStatusCompleted  = "completed"
	VideoStatusFailed     = "failed"
)
```

--------------------------------

### RealtimeUsage struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the structure for tracking token usage in real-time operations.

```go
type RealtimeUsage struct {
	TotalTokens        int                `json:"total_tokens"`
	InputTokens        int                `json:"input_tokens"`
	OutputTokens       int                `json:"output_tokens"`
	InputTokenDetails  InputTokenDetails  `json:"input_token_details"`
	OutputTokenDetails OutputTokenDetails `json:"output_token_details"`
}
```

--------------------------------

### GeminiUsageMetadata struct and merge function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Tracks token usage and billing metadata, with a utility to merge non-zero values.

```go
type GeminiUsageMetadata struct {
	PromptTokenCount           int                         `json:"promptTokenCount"`
	ToolUsePromptTokenCount    int                         `json:"toolUsePromptTokenCount"`
	CandidatesTokenCount       int                         `json:"candidatesTokenCount"`
	TotalTokenCount            int                         `json:"totalTokenCount"`
	ThoughtsTokenCount         int                         `json:"thoughtsTokenCount"`
	CachedContentTokenCount    int                         `json:"cachedContentTokenCount"`
	PromptTokensDetails        []GeminiPromptTokensDetails `json:"promptTokensDetails"`
	ToolUsePromptTokensDetails []GeminiPromptTokensDetails `json:"toolUsePromptTokensDetails"`
	CandidatesTokensDetails    []GeminiPromptTokensDetails `json:"candidatesTokensDetails"`
	BillingUsage               *BillingUsage               `json:"billing_usage,omitempty"`
}
```

```go
func MergeGeminiUsageMetadataNonZero(current *GeminiUsageMetadata, incoming *GeminiUsageMetadata) *GeminiUsageMetadata
```

--------------------------------

### GeneralOpenAIRequest GetSystemRoleName method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Method signature for retrieving the system role name from a GeneralOpenAIRequest.

```go
func (r *GeneralOpenAIRequest) GetSystemRoleName() string
```

--------------------------------

### RealtimeItem Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents an item within a real-time conversation.

```go
type RealtimeItem struct {
	Id        string            `json:"id"`
	Type      string            `json:"type"`
	Status    string            `json:"status"`
	Role      string            `json:"role"`
	Content   []RealtimeContent `json:"content"`
	Name      *string           `json:"name,omitempty"`
	ToolCalls any               `json:"tool_calls,omitempty"`
	CallId    string            `json:"call_id,omitempty"`
}
```

--------------------------------

### NewResponsesToChatStreamState constructor

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/internal/oai_responses

Constructor for ResponsesToChatStreamState. Takes a model string and includeUsage boolean.

```go
func NewResponsesToChatStreamState(model string, includeUsage bool) *ResponsesToChatStreamState
```

--------------------------------

### NewResponsesBufferedAccumulator constructor

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/internal/oai_responses

Constructor for ResponsesBufferedAccumulator. Returns a pointer to a new instance.

```go
func NewResponsesBufferedAccumulator() *ResponsesBufferedAccumulator
```

--------------------------------

### ResponsesBufferedAccumulator.BuildOutput method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/internal/oai_responses

Method on ResponsesBufferedAccumulator that builds the output as a slice of dto.ResponsesOutput.

```go
func (a *ResponsesBufferedAccumulator) BuildOutput() []dto.ResponsesOutput
```

--------------------------------

### ResponsesBufferedAccumulator.SupplementResponseOutput method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/internal/oai_responses

Method on ResponsesBufferedAccumulator that supplements the response output.

```go
func (a *ResponsesBufferedAccumulator) SupplementResponseOutput(resp *dto.OpenAIResponsesResponse)
```

--------------------------------

### AudioResponse struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the AudioResponse struct with a single Text field.

```go
type AudioResponse struct {
	Text string `json:"text"`
}
```

--------------------------------

### FromGemini function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Converts a Gemini chat request into a portable Intent. Use when extracting reasoning controls from a Gemini-format request.

```go
func FromGemini(req *dto.GeminiChatRequest) (Intent, error)
```

--------------------------------

### Mode type

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Defines the Mode type as a string.

```go
type Mode string
```

--------------------------------

### Notify struct and NewNotify constructor

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines a notification structure and its constructor function.

```go
type Notify struct {
	Type    string        `json:"type"`
	Title   string        `json:"title"`
	Content string        `json:"content"`
	Values  []interface{} `json:"values"`
}
```

```go
func NewNotify(t string, title string, content string, values []interface{}) Notify
```

--------------------------------

### ClaudeMessageSource Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents a source for Claude media content.

```go
type ClaudeMessageSource struct {
	Type      string `json:"type"`
	MediaType string `json:"media_type,omitempty"`
	Data      any    `json:"data,omitempty"`
	Url       string `json:"url,omitempty"`
}
```

--------------------------------

### ToolCallResponse struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents the response from a tool invocation.

```go
type ToolCallResponse struct {
	// Index is not nil only in chat completion chunk object
	Index    *int             `json:"index,omitempty"`
	ID       string           `json:"id,omitempty"`
	Type     any              `json:"type"`
	Function FunctionResponse `json:"function"`
}
```

--------------------------------

### Advanced Custom Path Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants identifying routes for model discovery and billing credit grants.

```go
const (
	// AdvancedCustomModelListPath identifies the optional OpenAI Models discovery route.
	AdvancedCustomModelListPath = "/v1/models"
	// AdvancedCustomBalancePath identifies the optional balance lookup route used by channel management.
	AdvancedCustomBalancePath = "/v1/dashboard/billing/credit_grants"
)
```

--------------------------------

### OpenAIVideoResponse Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the response structure for video file operations.

```go
type OpenAIVideoResponse struct {
	Id        string `json:"id" example:"file-abc123"`
	Object    string `json:"object" example:"file"`
	Bytes     int64  `json:"bytes" example:"120000"`
	CreatedAt int64  `json:"created_at" example:"1677610602"`
	ExpiresAt int64  `json:"expires_at" example:"1677614202"`
	Filename  string `json:"filename" example:"mydata.jsonl"`
	Purpose   string `json:"purpose" example:"fine-tune"`
}
```

--------------------------------

### OpenAIResponsesCompactionRequest struct and methods

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the request structure for OpenAI response compaction and its associated helper methods.

```go
type OpenAIResponsesCompactionRequest struct {
	Model              string          `json:"model"`
	Input              json.RawMessage `json:"input,omitempty"`
	Instructions       json.RawMessage `json:"instructions,omitempty"`
	PreviousResponseID string          `json:"previous_response_id,omitempty"`
	// Codex compact request parity:
	// https://github.com/openai/codex/commit/53d59722268dde82fb93c1f37964ce196c2a86d7
	// https://github.com/openai/codex/commit/5d6f23a27bf9c90709af527a7108c1c2eadf5123
	Tools                json.RawMessage `json:"tools,omitempty"`
	ParallelToolCalls    json.RawMessage `json:"parallel_tool_calls,omitempty"`
	Reasoning            *Reasoning      `json:"reasoning,omitempty"`
	ServiceTier          string          `json:"service_tier,omitempty"`
	PromptCacheKey       json.RawMessage `json:"prompt_cache_key,omitempty"`
	PromptCacheOptions   json.RawMessage `json:"prompt_cache_options,omitempty"`
	PromptCacheRetention json.RawMessage `json:"prompt_cache_retention,omitempty"`
	Text                 json.RawMessage `json:"text,omitempty"`
}
```

```go
func (r *OpenAIResponsesCompactionRequest) GetTokenCountMeta() *types.TokenCountMeta
```

```go
func (r *OpenAIResponsesCompactionRequest) IsStream(c *http.Request) bool
```

```go
func (r *OpenAIResponsesCompactionRequest) SetModelName(modelName string)
```

--------------------------------

### EffectiveEffort function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Returns the effective effort for the given intent.

```go
func EffectiveEffort(intent Intent) Effort
```

--------------------------------

### SetContentString method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Sets the content field to the given string. Use to update the delta content.

```go
func (c *ChatCompletionsStreamResponseChoiceDelta) SetContentString(s string)
```

--------------------------------

### ResponseFormat struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the format for API responses, including optional JSON schema.

```go
type ResponseFormat struct {
	Type       string          `json:"type,omitempty"`
	JsonSchema json.RawMessage `json:"json_schema,omitempty"`
}
```

--------------------------------

### Intent.IsEmpty method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Reports whether the intent is empty.

```go
func (i Intent) IsEmpty() bool
```

--------------------------------

### RealtimeResponse Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents a response in the real-time API.

```go
type RealtimeResponse struct {
	Usage *RealtimeUsage `json:"usage"`
}
```

--------------------------------

### Source constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Defines the possible Source constants: explicit, native, suffix, and pivot.

```go
const (
	SourceExplicit Source = "explicit"
	SourceNative   Source = "native"
	SourceSuffix   Source = "suffix"
	SourcePivot    Source = "pivot"
)
```

--------------------------------

### GeminiChatRequest SetTools

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

SetTools sets the tools for the request.

```go
func (r *GeminiChatRequest) SetTools(tools []GeminiChatTool)
```

--------------------------------

### NewOpenAIChatBillingUsage function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Creates a BillingUsage from a Usage for OpenAI chat.

```go
func NewOpenAIChatBillingUsage(usage *Usage) *BillingUsage
```

--------------------------------

### Response Output Type Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants for response output lifecycle events.

```go
const (
	ResponsesOutputTypeItemAdded = "response.output_item.added"
	ResponsesOutputTypeItemDone  = "response.output_item.done"
)
```

--------------------------------

### ClaudeRequest IsStringSystem

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Checks if the system prompt is a string.

```go
func (c *ClaudeRequest) IsStringSystem() bool
```

--------------------------------

### PlayGroundRequest Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Request structure for playground interactions.

```go
type PlayGroundRequest struct {
	Model string `json:"model,omitempty"`
	Group string `json:"group,omitempty"`
}
```

--------------------------------

### Realtime Event Type Constants (Output)

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants for outgoing realtime event types.

```go
const (
	RealtimeEventTypeResponseDone                   = "response.done"
	RealtimeEventTypeSessionUpdated                 = "session.updated"
	RealtimeEventTypeSessionCreated                 = "session.created"
	RealtimeEventResponseAudioDelta                 = "response.audio.delta"
	RealtimeEventResponseAudioTranscriptionDelta    = "response.audio_transcript.delta"
	RealtimeEventResponseFunctionCallArgumentsDelta = "response.function_call_arguments.delta"
	RealtimeEventResponseFunctionCallArgumentsDone  = "response.function_call_arguments.done"
	RealtimeEventConversationItemCreated            = "conversation.item.created"
)
```

--------------------------------

### Intent struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Defines the protocol-independent reasoning controls for a request. Summary visibility is intentionally independent from reasoning strength.

```go
type Intent struct {
	Mode            Mode
	Effort          Effort
	BudgetTokens    *int
	IncludeThoughts *bool
	Source          Source
	BudgetSource    Source
}
```

--------------------------------

### OpenAIModels struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents metadata for an OpenAI model.

```go
type OpenAIModels struct {
	Id                     string               `json:"id"`
	Object                 string               `json:"object"`
	Created                int                  `json:"created"`
	OwnedBy                string               `json:"owned_by"`
	SupportedEndpointTypes []types.EndpointType `json:"supported_endpoint_types"`
}
```

--------------------------------

### Advanced Custom Auth Type Constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constants defining authentication types for advanced custom configurations.

```go
const (
	AdvancedCustomAuthTypeNone   = "none"
	AdvancedCustomAuthTypeHeader = "header"
	AdvancedCustomAuthTypeQuery  = "query"
)
```

--------------------------------

### OpenAIResponsesRequest Struct Variables

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the fields for the OpenAIResponsesRequest struct, including optional fields for streaming, tool usage, and provider-specific configurations like Qwen and Perplexity.

```go
	PreviousResponseID string          `json:"previous_response_id,omitempty"`
	Reasoning          *Reasoning      `json:"reasoning,omitempty"`
	// ServiceTier specifies upstream service level and may affect billing.
	// This field is filtered by default and can be enabled via channel setting allow_service_tier.
	ServiceTier string `json:"service_tier,omitempty"`
	// Store controls whether upstream may store request/response data.
	// This field is allowed by default and can be disabled via channel setting disable_store.
	Store                json.RawMessage `json:"store,omitempty"`
	PromptCacheKey       json.RawMessage `json:"prompt_cache_key,omitempty"`
	PromptCacheOptions   json.RawMessage `json:"prompt_cache_options,omitempty"`
	PromptCacheRetention json.RawMessage `json:"prompt_cache_retention,omitempty"`
	// SafetyIdentifier carries client identity for policy abuse detection.
	// This field is filtered by default and can be enabled via channel setting allow_safety_identifier.
	SafetyIdentifier json.RawMessage `json:"safety_identifier,omitempty"`
	Stream           *bool           `json:"stream,omitempty"`
	StreamOptions    *StreamOptions  `json:"stream_options,omitempty"`
	Temperature      *float64        `json:"temperature,omitempty"`
	Text             json.RawMessage `json:"text,omitempty"`
	ToolChoice       json.RawMessage `json:"tool_choice,omitempty"`
	Tools            json.RawMessage `json:"tools,omitempty"` // 需要处理的参数很少，MCP 参数太多不确定，所以用 map
	TopP             *float64        `json:"top_p,omitempty"`
	Truncation       json.RawMessage `json:"truncation,omitempty"`
	User             json.RawMessage `json:"user,omitempty"`
	MaxToolCalls     *uint           `json:"max_tool_calls,omitempty"`
	Prompt           json.RawMessage `json:"prompt,omitempty"`
	// Codex Responses metadata/client_metadata:
	// https://github.com/openai/codex/commit/14df0e8833aad0d6d78287954b61ffac67af936c
	ClientMetadata json.RawMessage `json:"client_metadata,omitempty"`
	// qwen
	EnableThinking json.RawMessage `json:"enable_thinking,omitempty"`
	ThinkingBudget json.RawMessage `json:"thinking_budget,omitempty"`
	// perplexity
	Preset json.RawMessage `json:"preset,omitempty"`

	// Internal conversion state; never serialized to an upstream protocol.
	ReasoningConversion *ReasoningConversionState `json:"-"`
}
```

--------------------------------

### GeminiGroundingMetadata Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Contains metadata related to grounding, such as web search and retrieval queries.

```go
type GeminiGroundingMetadata struct {
	WebSearchQueries             []string        `json:"webSearchQueries,omitempty"`
	RetrievalQueries             []string        `json:"retrievalQueries,omitempty"`
	GroundingChunks              json.RawMessage `json:"groundingChunks,omitempty"`
	GroundingSupports            json.RawMessage `json:"groundingSupports,omitempty"`
	SearchEntryPoint             json.RawMessage `json:"searchEntryPoint,omitempty"`
	RetrievalMetadata            json.RawMessage `json:"retrievalMetadata,omitempty"`
	SourceFlaggingUris           json.RawMessage `json:"sourceFlaggingUris,omitempty"`
	GoogleMapsWidgetContextToken string          `json:"googleMapsWidgetContextToken,omitempty"`
}
```

--------------------------------

### VertexKeyType constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the supported key types for Vertex authentication.

```go
type VertexKeyType string
```

```go
const (
	VertexKeyTypeJSON   VertexKeyType = "json"
	VertexKeyTypeAPIKey VertexKeyType = "api_key"
)
```

--------------------------------

### SetReasoningContent method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Sets the reasoning content field to the given string. Use to update reasoning content.

```go
func (c *ChatCompletionsStreamResponseChoiceDelta) SetReasoningContent(s string)
```

--------------------------------

### GeneralOpenAIRequest fields with json.RawMessage and omitempty

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Field declarations for a request struct, showing the convention that unquoted parameters must use json.RawMessage with omitempty tags. Includes fields for search, Minimax, and vLLM options, plus an internal conversion state that is never serialized.

```go
SearchRecencyFilter    json.RawMessage `json:"search_recency_filter,omitempty"`
	ReturnImages           *bool           `json:"return_images,omitempty"`
	ReturnRelatedQuestions *bool           `json:"return_related_questions,omitempty"`
	SearchMode             json.RawMessage `json:"search_mode,omitempty"`
	// Minimax
	ReasoningSplit json.RawMessage `json:"reasoning_split,omitempty"`
	// vLLM
	ThinkingTokenBudget json.RawMessage `json:"thinking_token_budget,omitempty"`

	// Internal conversion state; never serialized to an upstream protocol.
	ReasoningConversion *ReasoningConversionState `json:"-"`
```

--------------------------------

### ResolveGeminiDefault function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Materializes documented family defaults when a conversion targets another protocol. Dynamic 2.5 defaults retain their -1 budget in the in-process pivot; Flash-Lite's default is explicitly off.

```go
func ResolveGeminiDefault(model string, intent Intent) Intent
```

--------------------------------

### GeminiPromptTokensDetails struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Details regarding token counts for specific modalities.

```go
type GeminiPromptTokensDetails struct {
	Modality   string `json:"modality"`
	TokenCount int    `json:"tokenCount"`
}
```

--------------------------------

### ShouldPreserveThinkingSuffix Method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/convmeta

Checks if the thinking suffix should be preserved for a given model name.

```go
func (o *Options) ShouldPreserveThinkingSuffix(modelName string) bool
```

--------------------------------

### OpenAIEffort function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Maps the canonical cross-provider vocabulary to the public OpenAI reasoning_effort vocabulary. Claude/OpenRouter "max" has no direct OpenAI equivalent and is represented by xhigh at that wire boundary.

```go
func OpenAIEffort(effort Effort) Effort
```

--------------------------------

### IntentFromState function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Reconstructs a portable intent from host- or pivot-carried conversion state. A nil state is an empty intent.

```go
func IntentFromState(state *dto.ReasoningConversionState) Intent
```

--------------------------------

### Request interface

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the interface for request handling, including token counting and model naming.

```go
type Request interface {
	GetTokenCountMeta() *types.TokenCountMeta
	IsStream(c *http.Request) bool
	SetModelName(modelName string)
}
```

--------------------------------

### OpenAIResponsesCompactionResponse struct and method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Defines the response structure for OpenAI response compaction and a method to retrieve errors.

```go
type OpenAIResponsesCompactionResponse struct {
	ID        string          `json:"id"`
	Object    string          `json:"object"`
	CreatedAt int             `json:"created_at"`
	Output    json.RawMessage `json:"output"`
	Usage     *Usage          `json:"usage"`
	Error     any             `json:"error,omitempty"`
}
```

```go
func (o *OpenAIResponsesCompactionResponse) GetOpenAIError() *types.OpenAIError
```

--------------------------------

### Custom Type Constant

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Constant for custom type identification.

```go
const CustomType = "custom"
```

--------------------------------

### GeminiModel Struct

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Represents the metadata for a Gemini model.

```go
type GeminiModel struct {
	Name                       interface{}   `json:"name"`
	BaseModelId                interface{}   `json:"baseModelId"`
	Version                    interface{}   `json:"version"`
	DisplayName                interface{}   `json:"displayName"`
	Description                interface{}   `json:"description"`
	InputTokenLimit            interface{}   `json:"inputTokenLimit"`
	OutputTokenLimit           interface{}   `json:"outputTokenLimit"`
```

--------------------------------

### Define ConversionLossPolicy type and constants

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/types

Defines policies for handling data loss during conversion, including safe, strict, and allow modes.

```go
type ConversionLossPolicy string
```

```go
const (
	// ConversionLossPolicySafe rejects request-phase conversions that would
	// change tool execution semantics, while returning non-fatal loss as
	// diagnostics. It is opt-in; the default is ConversionLossPolicyAllow.
	ConversionLossPolicySafe ConversionLossPolicy = "safe"
	// ConversionLossPolicyStrict rejects every lossy conversion, including
	// presentation-only metadata loss.
	ConversionLossPolicyStrict ConversionLossPolicy = "strict"
	// ConversionLossPolicyAllow is the default. It permits lossy conversion
	// and reports every loss through the conversion result.
	ConversionLossPolicyAllow ConversionLossPolicy = "allow"
)
```

--------------------------------

### ResponseStreamState.UsageText

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert

Returns a textual representation of the usage metadata.

```APIDOC
## ResponseStreamState.UsageText

### Description
Returns a text representation of the usage information.

### Method
func (s *ResponseStreamState) UsageText() string

### Returns
- **string** - Usage as formatted text.

```

--------------------------------

### ReasoningStateOf function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/convmeta

Nil-safe reader for Meta.ReasoningState. Added in v0.2.0.

```go
func ReasoningStateOf(m Meta) *dto.ReasoningConversionState
```

--------------------------------

### SetContent method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

Sets the content field to the given value. Use to update the message content.

```go
func (c *ClaudeMediaMessage) SetContent(content any)
```

--------------------------------

### NewResponsesToClaudeStreamState constructor

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/internal/oai_responses

Constructor for ResponsesToClaudeStreamState. Takes id and model strings. Added in v0.2.0.

```go
func NewResponsesToClaudeStreamState(id string, model string) *ResponsesToClaudeStreamState
```

--------------------------------

### ApplyToOpenAIChat function

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Writes the portable portion of an intent to the OpenAI pivot. reasoning_effort carries level-based strength; a JSON-excluded DTO state retains exact budgets and summary visibility across in-process steps.

```go
func ApplyToOpenAIChat(req *dto.GeneralOpenAIRequest, intent Intent) error
```

--------------------------------

### NewGeminiHostedStreamBridge

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert

Instantiates a GeminiHostedStreamBridge for Gemini hosted streaming responses.

```APIDOC
## NewGeminiHostedStreamBridge

### Description
Creates a new Gemini hosted stream bridge.

### Function
NewGeminiHostedStreamBridge() *GeminiHostedStreamBridge

### Returns
- **GeminiHostedStreamBridge** - Pointer to the new bridge.

```

--------------------------------

### EffectiveToolLossPolicy Method

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/convmeta

Returns the effective conversion loss policy for the options.

```go
func (o *Options) EffectiveToolLossPolicy() types.ConversionLossPolicy
```

--------------------------------

### OpenAI effort suffixes

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/relayconvert/reasoning

Defines the list of effort suffixes for OpenAI models.

```go
var OpenAIEffortSuffixes = []string{"-max", "-xhigh", "-high", "-medium", "-low", "-minimal", "-none"}
```

--------------------------------

### GeminiChatRequest GetTools

Source: https://pkg.go.dev/github.com/QuantumNous/new-api/relaykit/dto

GetTools returns the tools associated with the request.

```go
func (r *GeminiChatRequest) GetTools() []GeminiChatTool
```