客製化Providers

Extensions可以透過pi.registerProvider()註冊自訂模型提供者。這使得：

• 代理 - 透過公司代理或 API 網關路由請求
• 自訂端點 - 使用自架或私人模型部署
• OAuth/SSO - 為企業提供者新增身分驗證流程
• 自訂 APIs - 為非標準 LLM APIs 實現串流傳輸

請參閱這些完整的提供者範例：

• examples/extensions/custom-provider-anthropic/
• examples/extensions/custom-provider-gitlab-duo/

• Example Extensions
• Quick Reference
• Override Existing Provider
• Register New Provider
• Unregister Provider
• OAuth Support
• Custom Streaming API
• Context Overflow Errors
• Testing Your Implementation
• Config Reference
• Model Definition Reference

Extensions 可以註冊完整的 pi-ai Provider 或使用舊的提供者設定表單。當需要自訂身份驗證、過濾、刷新或流行為時，首選完整的提供者。 Pi 組成 models.json 覆蓋上面註冊的本地提供者。

import { createProvider, openAICompletionsApi } from "@earendil-works/pi-ai";
import type { ExtensionAPI } from "@earendil-works/pi-coding-agent";

export default function (pi: ExtensionAPI) {
  pi.registerProvider(createProvider({
    id: "native-local",
    name: "Native Local",
    baseUrl: "http://localhost:8080/v1",
    auth: {
      apiKey: {
        name: "Local server API key",
        async login(interaction) {
          return {
            type: "api_key",
            key: await interaction.prompt({ type: "secret", message: "API key" })
          };
        },
        async resolve({ credential }) {
          return credential?.key
            ? { auth: { apiKey: credential.key }, source: "stored API key" }
            : undefined;
        }
      }
    },
    models: [],
    api: openAICompletionsApi()
  }));

  // Legacy provider-config form:
  // Override baseUrl for existing provider
  pi.registerProvider("anthropic", {
    baseUrl: "https://proxy.example.com"
  });

  // Register new provider with models
  pi.registerProvider("my-provider", {
    name: "My Provider",
    baseUrl: "https://api.example.com",
    apiKey: "$MY_API_KEY",
    api: "openai-completions",
    models: [
      {
        id: "my-model",
        name: "My Model",
        reasoning: false,
        input: ["text", "image"],
        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
        contextWindow: 128000,
        maxTokens: 4096
      }
    ]
  });
}

擴充工廠也可以是async。對於動態模型發現，在工廠中取得並註冊模型而不是session_start。 pi 在啟動繼續之前等待工廠，因此提供程序在互動式啟動期間和 pi --list-models 期間可用。

最簡單的用例：透過代理程式重定向現有提供者。

// All Anthropic requests now go through your proxy
pi.registerProvider("anthropic", {
  baseUrl: "https://proxy.example.com"
});

// Add custom headers to OpenAI requests
pi.registerProvider("openai", {
  headers: {
    "X-Custom-Header": "value"
  }
});

// Both baseUrl and headers
pi.registerProvider("google", {
  baseUrl: "https://ai-gateway.corp.com/google",
  headers: {
    "X-Corp-Auth": "$CORP_AUTH_TOKEN"  // env var or literal
  }
});

當僅提供 baseUrl 和/或 headers（無 models）時，該提供者的所有現有模型都將與新端點一起保留。

若要新增全新的提供程序，請指定 models 以及所需的配置。

如果模型清單來自遠端端點，請使用非同步擴充工廠：

import type { ExtensionAPI } from "@earendil-works/pi-coding-agent";

export default async function (pi: ExtensionAPI) {
  const response = await fetch("http://localhost:1234/v1/models");
  const payload = (await response.json()) as {
    data: Array<{
      id: string;
      name?: string;
      context_window?: number;
      max_tokens?: number;
    }>;
  };

  pi.registerProvider("local-openai", {
    baseUrl: "http://localhost:1234/v1",
    apiKey: "$LOCAL_OPENAI_API_KEY",
    api: "openai-completions",
    models: payload.data.map((model) => ({
      id: model.id,
      name: model.name ?? model.id,
      reasoning: false,
      input: ["text"],
      cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
      contextWindow: model.context_window ?? 128000,
      maxTokens: model.max_tokens ?? 4096,
    })),
  });
}

這會在啟動完成之前註冊獲取的模型。

pi.registerProvider("my-llm", {
  baseUrl: "https://api.my-llm.com/v1",
  apiKey: "$MY_LLM_API_KEY",  // env var reference
  api: "openai-completions",  // which streaming API to use
  models: [
    {
      id: "my-llm-large",
      name: "My LLM Large",
      reasoning: true,        // supports extended thinking
      input: ["text", "image"],
      cost: {
        input: 3.0,           // $/million tokens
        output: 15.0,
        cacheRead: 0.3,
        cacheWrite: 3.75
      },
      contextWindow: 200000,
      maxTokens: 16384
    }
  ]
});

當提供 models 時，它會替換該提供者的所有現有模型。

apiKey 和自訂標頭值使用與 models.json 相同的設定值語法：開頭的 !command 會對整個值執行命令，$ENV_VAR 和 ${ENV_VAR} 會插入環境變數，$ 會輸出字面 $，$! 會輸出字面 !。

使用 pi.unregisterProvider(name) 刪除先前透過 pi.registerProvider(name,...) 註冊的提供者：

// Register
pi.registerProvider("my-llm", {
  baseUrl: "https://api.my-llm.com/v1",
  apiKey: "$MY_LLM_API_KEY",
  api: "openai-completions",
  models: [
    {
      id: "my-llm-large",
      name: "My LLM Large",
      reasoning: true,
      input: ["text", "image"],
      cost: { input: 3.0, output: 15.0, cacheRead: 0.3, cacheWrite: 3.75 },
      contextWindow: 200000,
      maxTokens: 16384
    }
  ]
});

// Later, remove it
pi.unregisterProvider("my-llm");

取消註冊會刪除該提供者的動態模型、API key 後備、OAuth 提供者註冊和自訂流程處理程序註冊。任何被覆蓋的內建模型或提供者行為都會被恢復。

初始擴充載入階段之後進行的呼叫會立即套用，因此不需要 /reload。

api字段決定使用哪種流實現：

API	用於
anthropic-messages	人擇克勞德 API 及其兼容者
openai-completions	OpenAI 聊天完成 API 和相容版本
openai-responses	OpenAI 回應 API
azure-openai-responses	Azure OpenAI 回應 API
openai-codex-responses	OpenAI Codex 回覆 API
mistral-conversations	本地米斯特拉爾聊天完成串流
google-generative-ai	谷歌生成人工智慧API
google-vertex	Google Vertex AI API
bedrock-converse-stream	亞馬遜 Bedrock 匡威 API

大多數與 OpenAI 相容的供應商都使用 openai-completions。使用模型級別 thinkingLevelMap 來實現特定於模型的思維級別，使用 compat 來實現提供者的怪癖。 xhigh 和 max 等級是可選的，需要非空映射條目，並且可能被不支援的孔分隔：

models: [{
  id: "custom-model",
  // ...
  reasoning: true,
  thinkingLevelMap: {              // map pi levels to provider values; null hides unsupported levels
    minimal: null,
    low: null,
    medium: null,
    high: "default",
    xhigh: null,
    max: "max"
  },
  compat: {
    supportsDeveloperRole: false,   // use "system" instead of "developer"
    supportsReasoningEffort: true,
    maxTokensField: "max_tokens",   // instead of "max_completion_tokens"
    requiresToolResultName: true,   // tool results need name field
    thinkingFormat: "qwen",        // top-level enable_thinking: true
    cacheControlFormat: "anthropic" // Anthropic-style cache_control markers
  }
}]

將 openrouter 用於 OpenRouter 樣式 reasoning: { effort } 控制。將 together 用於 Together 樣式 reasoning: { enabled } 控制；對於supportsReasoningEffort，它也發送reasoning_effort。對於讀取 chat_template_kwargs.enable_thinking 並需要 preserve_thinking 的本機 Qwen 相容伺服器，請使用 qwen-chat-template。 將 cacheControlFormat: "anthropic" 用於與 OpenAI 相容的提供程序，透過 cache_control 在系統提示、最後一個工具定義以及最後一個使用者、助手或工具結果文字內容上公開人類風格的提示快取。

對於使用api: "anthropic-messages"的人類相容提供者，在其上游模型需要自適應思維的模型或提供者上設定compat.forceAdaptiveThinking: true（thinking.type: "adaptive"加output_config.effort）。內建自適應克勞德模型會自動設定此功能。僅針對發出空思維簽名並期望重播時 signature: "" 的提供者設定 compat.allowEmptySignature: true。

遷移注意：米斯特拉爾從openai-completions移至mistral-conversations。 對原生 Mistral 模型使用 mistral-conversations。 如果您有意透過 openai-completions 路由 Mistral 相容/自訂端點，請根據需要明確設定 compat 標誌。

如果您的提供者期望 Authorization: Bearer <key> 但不使用標準 API，請設定 authHeader: true：

pi.registerProvider("custom-api", {
  baseUrl: "https://api.example.com",
  apiKey: "$MY_API_KEY",
  authHeader: true,  // adds Authorization: Bearer header
  api: "openai-completions",
  models: [...]
});

每個請求都會解析密鑰。明確請求 Authorization 標頭優先於產生的值。

新增與/login整合的OAuth/SSO身份驗證：

import type { OAuthCredentials, OAuthLoginCallbacks } from "@earendil-works/pi-ai";

pi.registerProvider("corporate-ai", {
  baseUrl: "https://ai.corp.com/v1",
  api: "openai-responses",
  models: [...],
  oauth: {
    name: "Corporate AI (SSO)",

    async login(callbacks: OAuthLoginCallbacks): Promise<OAuthCredentials> {
      const method = await callbacks.onSelect({
        message: "Select login method:",
        options: [
          { id: "browser", label: "Browser OAuth" },
          { id: "device", label: "Device code" }
        ]
      });
      if (!method) throw new Error("Login cancelled");

      let code: string;
      if (method === "device") {
        callbacks.onDeviceCode({
          userCode: "ABCD-1234",
          verificationUri: "https://sso.corp.com/device",
          intervalSeconds: 5,
          expiresInSeconds: 900
        });
        code = await pollDeviceCodeUntilComplete();
      } else {
        callbacks.onAuth({ url: "https://sso.corp.com/authorize?..." });
        code = await callbacks.onPrompt({ message: "Enter SSO code:" });
      }

      // Exchange for tokens (your implementation)
      const tokens = await exchangeCodeForTokens(code);

      return {
        refresh: tokens.refreshToken,
        access: tokens.accessToken,
        expires: Date.now() + tokens.expiresIn * 1000
      };
    },

    async refreshToken(credentials: OAuthCredentials, signal: AbortSignal): Promise<OAuthCredentials> {
      const tokens = await refreshAccessToken(credentials.refresh, signal);
      return {
        refresh: tokens.refreshToken ?? credentials.refresh,
        access: tokens.accessToken,
        expires: Date.now() + tokens.expiresIn * 1000
      };
    },

    getApiKey(credentials: OAuthCredentials): string {
      return credentials.access;
    }
  }
});

註冊後，用戶可以透過/login corporate-ai進行身份驗證。

callbacks 物件為提供者擁有的流程提供 UI 中立的互動：

interface OAuthLoginCallbacks {
  // Open URL in browser (for OAuth redirects)
  onAuth(params: { url: string }): void;

  // Show device code (for device authorization flow)
  onDeviceCode(params: {
    userCode: string;
    verificationUri: string;
    intervalSeconds?: number;
    expiresInSeconds?: number;
  }): void;

  // Show transient progress
  onProgress?(message: string): void;

  // Prompt user for input (for manual token entry)
  onPrompt(params: { message: string }): Promise<string>;

  // Show an interactive selector, e.g. to choose browser OAuth vs device code
  onSelect(params: {
    message: string;
    options: { id: string; label: string }[];
  }): Promise<string | undefined>;
}

憑證保存在 ~/.pi/agent/auth.json 中：

interface OAuthCredentials {
  refresh: string;   // Refresh token (for refreshToken())
  access: string;    // Access token (returned by getApiKey())
  expires: number;   // Expiration timestamp in milliseconds
}

對於具有非標準API的供應商，實作streamSimple。在編寫自己的提供程序之前，請先研究現有的提供者實作：

參考實作：

• anthropic.ts - 人為訊息 API
• mistral.ts - 米斯特拉爾對話 API
• openai-completions.ts - OpenAI 聊天完成
• openai-responses.ts - OpenAI 回應 API
• google.ts - 谷歌生成式人工智慧
• amazon-bedrock.ts - AWS 基岩

所有提供者都遵循相同的模式：

import {
  type AssistantMessage,
  type AssistantMessageEventStream,
  type Context,
  type Model,
  type SimpleStreamOptions,
  calculateCost,
  createAssistantMessageEventStream,
} from "@earendil-works/pi-ai";

function streamMyProvider(
  model: Model<any>,
  context: Context,
  options?: SimpleStreamOptions
): AssistantMessageEventStream {
  const stream = createAssistantMessageEventStream();

  (async () => {
    // Initialize output message
    const output: AssistantMessage = {
      role: "assistant",
      content: [],
      api: model.api,
      provider: model.provider,
      model: model.id,
      usage: {
        input: 0,
        output: 0,
        cacheRead: 0,
        cacheWrite: 0,
        totalTokens: 0,
        cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },
      },
      stopReason: "pending",
      timestamp: Date.now(),
    };

    try {
      // Push start event
      stream.push({ type: "start", partial: output });

      // Make API request and process response...
      // Push content events as they arrive and set stopReason from the terminal event.
      if (output.stopReason === "pending") {
        throw new Error("Provider stream ended without a stop reason");
      }
      if (output.stopReason === "error" || output.stopReason === "aborted") {
        throw new Error(output.errorMessage || "An unknown error occurred");
      }

      // Push done event
      stream.push({
        type: "done",
        reason: output.stopReason,
        message: output
      });
      stream.end();
    } catch (error) {
      output.stopReason = options?.signal?.aborted ? "aborted" : "error";
      output.errorMessage = error instanceof Error ? error.message : String(error);
      stream.push({ type: "error", reason: output.stopReason, error: output });
      stream.end();
    }
  })();

  return stream;
}

依以下順序透過 stream.push() 推送事件：

1. { type: "start", partial: output } - 直播開始

2. 內容事件（可重複，追蹤每個區塊的contentIndex）：

   • { type: "text_start", contentIndex, partial } - 文字區塊開始
   • { type: "text_delta", contentIndex, delta, partial } - 文字區塊
   • { type: "text_end", contentIndex, content, partial } - 文字區塊結束
   • { type: "thinking_start", contentIndex, partial } - 思考開始
   • { type: "thinking_delta", contentIndex, delta, partial } - 思考塊
   • { type: "thinking_end", contentIndex, content, partial } - 思考結束
   • { type: "toolcall_start", contentIndex, partial } - 工具呼叫開始
   • { type: "toolcall_delta", contentIndex, delta, partial } - 工具呼叫JSON塊
   • { type: "toolcall_end", contentIndex, toolCall, partial } - 工具呼叫結束

3. { type: "done", reason, message } 或 { type: "error", reason, error } - 直播結束

每個事件中的 partial 欄位包含目前 AssistantMessage 狀態。收到資料時更新 output.content，然後將 output 包含為 partial。

當內容塊到達時將其加到 output.content：

// Text block
output.content.push({ type: "text", text: "" });
stream.push({ type: "text_start", contentIndex: output.content.length - 1, partial: output });

// As text arrives
const block = output.content[contentIndex];
if (block.type === "text") {
  block.text += delta;
  stream.push({ type: "text_delta", contentIndex, delta, partial: output });
}

// When block completes
stream.push({ type: "text_end", contentIndex, content: block.text, partial: output });

工具呼叫需要累加JSON並解析：

// Start tool call
output.content.push({
  type: "toolCall",
  id: toolCallId,
  name: toolName,
  arguments: {}
});
stream.push({ type: "toolcall_start", contentIndex: output.content.length - 1, partial: output });

// Accumulate JSON
let partialJson = "";
partialJson += jsonDelta;
try {
  block.arguments = JSON.parse(partialJson);
} catch {}
stream.push({ type: "toolcall_delta", contentIndex, delta: jsonDelta, partial: output });

// Complete
stream.push({
  type: "toolcall_end",
  contentIndex,
  toolCall: { type: "toolCall", id, name, arguments: block.arguments },
  partial: output
});

從 API 回應更新使用情況並計算成本：

output.usage.input = response.usage.input_tokens;
output.usage.output = response.usage.output_tokens;
output.usage.cacheRead = response.usage.cache_read_tokens ?? 0;
output.usage.cacheWrite = response.usage.cache_write_tokens ?? 0;
output.usage.totalTokens = output.usage.input + output.usage.output +
                           output.usage.cacheRead + output.usage.cacheWrite;
calculateCost(model, output.usage);

當請求超出模型的上下文視窗時，pi 可以透過壓縮對話並重試來自動恢復。只有當 pi 將故障識別為溢出時，此恢復才會啟動。

檢測在最終確定的助理訊息上運行：

• stopReason === "error"
• errorMessage 匹配 pi 的已知溢出模式之一（參見 packages/ai/src/utils/overflow.ts）

如果您的提供者傳回溢位錯誤並顯示 pi 無法識別的訊息，請規範化來自註冊​​提供者的相同擴充功能的錯誤。使用 message_end 處理程序重寫助手訊息，使其 errorMessage 以 pi 識別的短語開頭。通用後備context_length_exceeded是最安全的選擇。

const MY_PROVIDER_OVERFLOW_PATTERN = /your provider's overflow phrase/i;

export default function (pi: ExtensionAPI) {
  pi.registerProvider("my-provider", { /* ... */ });

  pi.on("message_end", (event, ctx) => {
    const message = event.message;
    if (message.role !== "assistant") return;
    if (message.stopReason !== "error") return;
    if (
      message.provider !== "my-provider" &&
      ctx.model?.provider !== "my-provider"
    )
      return;

    const errorMessage = message.errorMessage ?? "";
    if (errorMessage.includes("context_length_exceeded")) return;
    if (!MY_PROVIDER_OVERFLOW_PATTERN.test(errorMessage)) return;

    return {
      message: {
        ...message,
        errorMessage: `context_length_exceeded: ${errorMessage}`,
      },
    };
  });
}

message_end在pi追蹤自動壓縮的輔助訊息之前運行，因此重寫的errorMessage是pi檢查的內容。完成此操作後，pi 將：

1. 檢測從errorMessage開始的溢出。
2. 從即時上下文中刪除失敗的助手訊息。
3. 運轉壓實。
4. 重試該請求一次。

仔細保護重寫：

• 將其範圍限定為您的提供者（message.provider 和 ctx.model?.provider），因此來自其他提供者的不相關錯誤不會受到影響。
• 匹配特定於提供者的模式，而不是 pi 的通用溢出模式。重寫速率限製或限制錯誤（rate limit、too many requests）會錯誤地觸發壓縮，而不是 pi 的正常重試與回退路徑。
• 當 errorMessage 已包含 context_length_exceeded 時跳過，因此處理程序是冪等的。

註冊您的流函數：

pi.registerProvider("my-provider", {
  baseUrl: "https://api.example.com",
  apiKey: "$MY_API_KEY",
  api: "my-custom-api",
  models: [...],
  streamSimple: streamMyProvider
});

根據內建提供者使用的相同測試套件來測試您的提供者。從 packages/ai/test/ 複製並調整這些測試檔案：

測試	目的
stream.test.ts	基本串流、文字輸出
tokens.test.ts	令牌計數和使用
abort.test.ts	Abort訊號處理
empty.test.ts	空/最少回复
context-overflow.test.ts	上下文視窗限制
image-limits.test.ts	影像輸入處理
unicode-surrogate.test.ts	Unicode 邊緣狀況
tool-call-without-result.test.ts	工具呼叫邊緣情況
image-tool-result.test.ts	工具結果中的影像
total-tokens.test.ts	總代幣計算
cross-provider-handoff.test.ts	提供者之間的上下文切換

使用您的提供者/模型對執行測試以驗證相容性。

interface ProviderConfig {
  /** Display name for the provider in UI such as /login. */
  name?: string;

  /** API endpoint URL. Required when defining models. */
  baseUrl?: string;

  /** API key literal, env interpolation ($ENV_VAR or ${ENV_VAR}), or !command. Required when defining models (unless oauth). */
  apiKey?: string;

  /** API type for streaming. Required at provider or model level when defining models. */
  api?: Api;

  /** Custom streaming implementation for non-standard APIs. */
  streamSimple?: (
    model: Model<Api>,
    context: Context,
    options?: SimpleStreamOptions
  ) => AssistantMessageEventStream;

  /** Custom headers to include in requests. Values use the same resolution syntax as apiKey. */
  headers?: Record<string, string>;

  /** If true, adds Authorization: Bearer header with the resolved API key. */
  authHeader?: boolean;

  /** Models to register. If provided, replaces all existing models for this provider. */
  models?: ProviderModelConfig[];

  /** OAuth provider for /login support. */
  oauth?: {
    name: string;
    login(callbacks: OAuthLoginCallbacks): Promise<OAuthCredentials>;
    refreshToken(credentials: OAuthCredentials, signal: AbortSignal): Promise<OAuthCredentials>;
    getApiKey(credentials: OAuthCredentials): string;
  };
}

interface ProviderModelConfig {
  /** Model ID (e.g., "claude-sonnet-4-20250514"). */
  id: string;

  /** Display name (e.g., "Claude 4 Sonnet"). */
  name: string;

  /** API type override for this specific model. */
  api?: Api;

  /** API endpoint URL override for this specific model. */
  baseUrl?: string;

  /** Whether the model supports extended thinking. */
  reasoning: boolean;

  /** Maps pi thinking levels to provider/model-specific values; null marks a level unsupported. */
  thinkingLevelMap?: Partial<Record<"off" | "minimal" | "low" | "medium" | "high" | "xhigh" | "max", string | null>>;

  /** Supported input types. */
  input: ("text" | "image")[];

  /** Cost per million tokens (for usage tracking). */
  cost: {
    input: number;
    output: number;
    cacheRead: number;
    cacheWrite: number;
  };

  /** Maximum context window size in tokens. */
  contextWindow: number;

  /** Maximum output tokens. */
  maxTokens: number;

  /** Custom headers for this specific model. */
  headers?: Record<string, string>;

  /** Compatibility settings for the selected API. */
  compat?: {
    // openai-completions
    supportsStore?: boolean;
    supportsDeveloperRole?: boolean;
    supportsReasoningEffort?: boolean;
    supportsUsageInStreaming?: boolean;
    supportsFinishReason?: boolean;
    supportsStrictMode?: boolean;
    supportsOpenAIGrammarTools?: boolean; // openai-completions/openai-responses; false falls back to normal function tools
    maxTokensField?: "max_completion_tokens" | "max_tokens";
    requiresToolResultName?: boolean;
    requiresAssistantAfterToolResult?: boolean;
    requiresThinkingAsText?: boolean;
    requiresReasoningContentOnAssistantMessages?: boolean;
    thinkingFormat?: "openai" | "openrouter" | "deepseek" | "together" | "baseten" | "zai" | "qwen" | "chat-template" | "qwen-chat-template" | "string-thinking" | "ant-ling";
    chatTemplateKwargs?: Record<string, string | number | boolean | null | { "$var": "thinking.enabled" | "thinking.effort"; omitWhenOff?: boolean }>;
    chatTemplateArgs?: Record<string, string | number | boolean | null | { "$var": "thinking.enabled" | "thinking.effort"; omitWhenOff?: boolean }>;
    cacheControlFormat?: "anthropic";
    sessionAffinityFormat?: "openai" | "openai-nosession" | "openrouter";
    sendSessionAffinityHeaders?: boolean;

    // anthropic-messages
    supportsEagerToolInputStreaming?: boolean;
    supportsLongCacheRetention?: boolean;
    sendSessionAffinityHeaders?: boolean;
    supportsCacheControlOnTools?: boolean;
    forceAdaptiveThinking?: boolean;
    allowEmptySignature?: boolean;
    supportsStrictTools?: boolean;
  };
}

openrouter 發送reasoning: { effort }。啟用後，deepseek 會發送 thinking: { type: "enabled" | "disabled" } 和 reasoning_effort。當supportsReasoningEffort啟用時，together會發送reasoning: { enabled }，也會發送reasoning_effort。 qwen 適用於 DashScope 樣式的頂級 enable_thinking。對於讀取 chat_template_kwargs.enable_thinking 且需要 preserve_thinking 的本機 Qwen 相容伺服器，請使用 qwen-chat-template。使用 chat-template 來配置 chat_template_kwargs，例如 vLLM 後面的 DeepSeek V3.x 帶有 chatTemplateKwargs: { "thinking": { "$var": "thinking.enabled" } }。當提供者期望切換值低於 chat_template_args 並可選擇支援頂級 reasoning_effort 時，請使用 thinkingFormat: "baseten" 和 chatTemplateArgs。 cacheControlFormat: "anthropic" 將人類風格的 cache_control 標記應用於系統提示、最後一個工具定義以及最後一個使用者、助手或工具結果文字內容。