@workers-community/workers-types 4.20251120.0 → 4.20251125.0

package/index.d.ts

@@ -4109,6 +4109,427 @@ declare abstract class BaseAiTranslation {
   inputs: AiTranslationInput;
   postProcessedOutputs: AiTranslationOutput;
 }
+/**
+ * Workers AI support for OpenAI's Responses API
+ * Reference: https://github.com/openai/openai-node/blob/master/src/resources/responses/responses.ts
+ *
+ * It's a stripped down version from its source.
+ * It currently supports basic function calling, json mode and accepts images as input.
+ *
+ * It does not include types for WebSearch, CodeInterpreter, FileInputs, MCP, CustomTools.
+ * We plan to add those incrementally as model + platform capabilities evolve.
+ */
+type ResponsesInput = {
+  background?: boolean | null;
+  conversation?: string | ResponseConversationParam | null;
+  include?: Array<ResponseIncludable> | null;
+  input?: string | ResponseInput;
+  instructions?: string | null;
+  max_output_tokens?: number | null;
+  parallel_tool_calls?: boolean | null;
+  previous_response_id?: string | null;
+  prompt_cache_key?: string;
+  reasoning?: Reasoning | null;
+  safety_identifier?: string;
+  service_tier?: "auto" | "default" | "flex" | "scale" | "priority" | null;
+  stream?: boolean | null;
+  stream_options?: StreamOptions | null;
+  temperature?: number | null;
+  text?: ResponseTextConfig;
+  tool_choice?: ToolChoiceOptions | ToolChoiceFunction;
+  tools?: Array<Tool>;
+  top_p?: number | null;
+  truncation?: "auto" | "disabled" | null;
+};
+type ResponsesOutput = {
+  id?: string;
+  created_at?: number;
+  output_text?: string;
+  error?: ResponseError | null;
+  incomplete_details?: ResponseIncompleteDetails | null;
+  instructions?: string | Array<ResponseInputItem> | null;
+  object?: "response";
+  output?: Array<ResponseOutputItem>;
+  parallel_tool_calls?: boolean;
+  temperature?: number | null;
+  tool_choice?: ToolChoiceOptions | ToolChoiceFunction;
+  tools?: Array<Tool>;
+  top_p?: number | null;
+  max_output_tokens?: number | null;
+  previous_response_id?: string | null;
+  prompt?: ResponsePrompt | null;
+  reasoning?: Reasoning | null;
+  safety_identifier?: string;
+  service_tier?: "auto" | "default" | "flex" | "scale" | "priority" | null;
+  status?: ResponseStatus;
+  text?: ResponseTextConfig;
+  truncation?: "auto" | "disabled" | null;
+  usage?: ResponseUsage;
+};
+type EasyInputMessage = {
+  content: string | ResponseInputMessageContentList;
+  role: "user" | "assistant" | "system" | "developer";
+  type?: "message";
+};
+type ResponsesFunctionTool = {
+  name: string;
+  parameters: {
+    [key: string]: unknown;
+  } | null;
+  strict: boolean | null;
+  type: "function";
+  description?: string | null;
+};
+type ResponseIncompleteDetails = {
+  reason?: "max_output_tokens" | "content_filter";
+};
+type ResponsePrompt = {
+  id: string;
+  variables?: {
+    [key: string]: string | ResponseInputText | ResponseInputImage;
+  } | null;
+  version?: string | null;
+};
+type Reasoning = {
+  effort?: ReasoningEffort | null;
+  generate_summary?: "auto" | "concise" | "detailed" | null;
+  summary?: "auto" | "concise" | "detailed" | null;
+};
+type ResponseContent =
+  | ResponseInputText
+  | ResponseInputImage
+  | ResponseOutputText
+  | ResponseOutputRefusal
+  | ResponseContentReasoningText;
+type ResponseContentReasoningText = {
+  text: string;
+  type: "reasoning_text";
+};
+type ResponseConversationParam = {
+  id: string;
+};
+type ResponseCreatedEvent = {
+  response: Response;
+  sequence_number: number;
+  type: "response.created";
+};
+type ResponseCustomToolCallOutput = {
+  call_id: string;
+  output: string | Array<ResponseInputText | ResponseInputImage>;
+  type: "custom_tool_call_output";
+  id?: string;
+};
+type ResponseError = {
+  code:
+    | "server_error"
+    | "rate_limit_exceeded"
+    | "invalid_prompt"
+    | "vector_store_timeout"
+    | "invalid_image"
+    | "invalid_image_format"
+    | "invalid_base64_image"
+    | "invalid_image_url"
+    | "image_too_large"
+    | "image_too_small"
+    | "image_parse_error"
+    | "image_content_policy_violation"
+    | "invalid_image_mode"
+    | "image_file_too_large"
+    | "unsupported_image_media_type"
+    | "empty_image_file"
+    | "failed_to_download_image"
+    | "image_file_not_found";
+  message: string;
+};
+type ResponseErrorEvent = {
+  code: string | null;
+  message: string;
+  param: string | null;
+  sequence_number: number;
+  type: "error";
+};
+type ResponseFailedEvent = {
+  response: Response;
+  sequence_number: number;
+  type: "response.failed";
+};
+type ResponseFormatText = {
+  type: "text";
+};
+type ResponseFormatJSONObject = {
+  type: "json_object";
+};
+type ResponseFormatTextConfig =
+  | ResponseFormatText
+  | ResponseFormatTextJSONSchemaConfig
+  | ResponseFormatJSONObject;
+type ResponseFormatTextJSONSchemaConfig = {
+  name: string;
+  schema: {
+    [key: string]: unknown;
+  };
+  type: "json_schema";
+  description?: string;
+  strict?: boolean | null;
+};
+type ResponseFunctionCallArgumentsDeltaEvent = {
+  delta: string;
+  item_id: string;
+  output_index: number;
+  sequence_number: number;
+  type: "response.function_call_arguments.delta";
+};
+type ResponseFunctionCallArgumentsDoneEvent = {
+  arguments: string;
+  item_id: string;
+  name: string;
+  output_index: number;
+  sequence_number: number;
+  type: "response.function_call_arguments.done";
+};
+type ResponseFunctionCallOutputItem =
+  | ResponseInputTextContent
+  | ResponseInputImageContent;
+type ResponseFunctionCallOutputItemList = Array<ResponseFunctionCallOutputItem>;
+type ResponseFunctionToolCall = {
+  arguments: string;
+  call_id: string;
+  name: string;
+  type: "function_call";
+  id?: string;
+  status?: "in_progress" | "completed" | "incomplete";
+};
+interface ResponseFunctionToolCallItem extends ResponseFunctionToolCall {
+  id: string;
+}
+type ResponseFunctionToolCallOutputItem = {
+  id: string;
+  call_id: string;
+  output: string | Array<ResponseInputText | ResponseInputImage>;
+  type: "function_call_output";
+  status?: "in_progress" | "completed" | "incomplete";
+};
+type ResponseIncludable =
+  | "message.input_image.image_url"
+  | "message.output_text.logprobs";
+type ResponseIncompleteEvent = {
+  response: Response;
+  sequence_number: number;
+  type: "response.incomplete";
+};
+type ResponseInput = Array<ResponseInputItem>;
+type ResponseInputContent = ResponseInputText | ResponseInputImage;
+type ResponseInputImage = {
+  detail: "low" | "high" | "auto";
+  type: "input_image";
+  /**
+   * Base64 encoded image
+   */
+  image_url?: string | null;
+};
+type ResponseInputImageContent = {
+  type: "input_image";
+  detail?: "low" | "high" | "auto" | null;
+  /**
+   * Base64 encoded image
+   */
+  image_url?: string | null;
+};
+type ResponseInputItem =
+  | EasyInputMessage
+  | ResponseInputItemMessage
+  | ResponseOutputMessage
+  | ResponseFunctionToolCall
+  | ResponseInputItemFunctionCallOutput
+  | ResponseReasoningItem;
+type ResponseInputItemFunctionCallOutput = {
+  call_id: string;
+  output: string | ResponseFunctionCallOutputItemList;
+  type: "function_call_output";
+  id?: string | null;
+  status?: "in_progress" | "completed" | "incomplete" | null;
+};
+type ResponseInputItemMessage = {
+  content: ResponseInputMessageContentList;
+  role: "user" | "system" | "developer";
+  status?: "in_progress" | "completed" | "incomplete";
+  type?: "message";
+};
+type ResponseInputMessageContentList = Array<ResponseInputContent>;
+type ResponseInputMessageItem = {
+  id: string;
+  content: ResponseInputMessageContentList;
+  role: "user" | "system" | "developer";
+  status?: "in_progress" | "completed" | "incomplete";
+  type?: "message";
+};
+type ResponseInputText = {
+  text: string;
+  type: "input_text";
+};
+type ResponseInputTextContent = {
+  text: string;
+  type: "input_text";
+};
+type ResponseItem =
+  | ResponseInputMessageItem
+  | ResponseOutputMessage
+  | ResponseFunctionToolCallItem
+  | ResponseFunctionToolCallOutputItem;
+type ResponseOutputItem =
+  | ResponseOutputMessage
+  | ResponseFunctionToolCall
+  | ResponseReasoningItem;
+type ResponseOutputItemAddedEvent = {
+  item: ResponseOutputItem;
+  output_index: number;
+  sequence_number: number;
+  type: "response.output_item.added";
+};
+type ResponseOutputItemDoneEvent = {
+  item: ResponseOutputItem;
+  output_index: number;
+  sequence_number: number;
+  type: "response.output_item.done";
+};
+type ResponseOutputMessage = {
+  id: string;
+  content: Array<ResponseOutputText | ResponseOutputRefusal>;
+  role: "assistant";
+  status: "in_progress" | "completed" | "incomplete";
+  type: "message";
+};
+type ResponseOutputRefusal = {
+  refusal: string;
+  type: "refusal";
+};
+type ResponseOutputText = {
+  text: string;
+  type: "output_text";
+  logprobs?: Array<Logprob>;
+};
+type ResponseReasoningItem = {
+  id: string;
+  summary: Array<ResponseReasoningSummaryItem>;
+  type: "reasoning";
+  content?: Array<ResponseReasoningContentItem>;
+  encrypted_content?: string | null;
+  status?: "in_progress" | "completed" | "incomplete";
+};
+type ResponseReasoningSummaryItem = {
+  text: string;
+  type: "summary_text";
+};
+type ResponseReasoningContentItem = {
+  text: string;
+  type: "reasoning_text";
+};
+type ResponseReasoningTextDeltaEvent = {
+  content_index: number;
+  delta: string;
+  item_id: string;
+  output_index: number;
+  sequence_number: number;
+  type: "response.reasoning_text.delta";
+};
+type ResponseReasoningTextDoneEvent = {
+  content_index: number;
+  item_id: string;
+  output_index: number;
+  sequence_number: number;
+  text: string;
+  type: "response.reasoning_text.done";
+};
+type ResponseRefusalDeltaEvent = {
+  content_index: number;
+  delta: string;
+  item_id: string;
+  output_index: number;
+  sequence_number: number;
+  type: "response.refusal.delta";
+};
+type ResponseRefusalDoneEvent = {
+  content_index: number;
+  item_id: string;
+  output_index: number;
+  refusal: string;
+  sequence_number: number;
+  type: "response.refusal.done";
+};
+type ResponseStatus =
+  | "completed"
+  | "failed"
+  | "in_progress"
+  | "cancelled"
+  | "queued"
+  | "incomplete";
+type ResponseStreamEvent =
+  | ResponseCompletedEvent
+  | ResponseCreatedEvent
+  | ResponseErrorEvent
+  | ResponseFunctionCallArgumentsDeltaEvent
+  | ResponseFunctionCallArgumentsDoneEvent
+  | ResponseFailedEvent
+  | ResponseIncompleteEvent
+  | ResponseOutputItemAddedEvent
+  | ResponseOutputItemDoneEvent
+  | ResponseReasoningTextDeltaEvent
+  | ResponseReasoningTextDoneEvent
+  | ResponseRefusalDeltaEvent
+  | ResponseRefusalDoneEvent
+  | ResponseTextDeltaEvent
+  | ResponseTextDoneEvent;
+type ResponseCompletedEvent = {
+  response: Response;
+  sequence_number: number;
+  type: "response.completed";
+};
+type ResponseTextConfig = {
+  format?: ResponseFormatTextConfig;
+  verbosity?: "low" | "medium" | "high" | null;
+};
+type ResponseTextDeltaEvent = {
+  content_index: number;
+  delta: string;
+  item_id: string;
+  logprobs: Array<Logprob>;
+  output_index: number;
+  sequence_number: number;
+  type: "response.output_text.delta";
+};
+type ResponseTextDoneEvent = {
+  content_index: number;
+  item_id: string;
+  logprobs: Array<Logprob>;
+  output_index: number;
+  sequence_number: number;
+  text: string;
+  type: "response.output_text.done";
+};
+type Logprob = {
+  token: string;
+  logprob: number;
+  top_logprobs?: Array<TopLogprob>;
+};
+type TopLogprob = {
+  token?: string;
+  logprob?: number;
+};
+type ResponseUsage = {
+  input_tokens: number;
+  output_tokens: number;
+  total_tokens: number;
+};
+type Tool = ResponsesFunctionTool;
+type ToolChoiceFunction = {
+  name: string;
+  type: "function";
+};
+type ToolChoiceOptions = "none";
+type ReasoningEffort = "minimal" | "low" | "medium" | "high" | null;
+type StreamOptions = {
+  include_obfuscation?: boolean;
+};
 type Ai_Cf_Baai_Bge_Base_En_V1_5_Input =
   | {
       text: string | string[];
@@ -4141,8 +4562,8 @@ type Ai_Cf_Baai_Bge_Base_En_V1_5_Output =
        */
       pooling?: "mean" | "cls";
     }
-  | AsyncResponse;
-interface AsyncResponse {
+  | Ai_Cf_Baai_Bge_Base_En_V1_5_AsyncResponse;
+interface Ai_Cf_Baai_Bge_Base_En_V1_5_AsyncResponse {
   /**
    * The async request id that can be used to obtain the results.
    */
@@ -4224,7 +4645,13 @@ type Ai_Cf_Meta_M2M100_1_2B_Output =
        */
       translated_text?: string;
     }
-  | AsyncResponse;
+  | Ai_Cf_Meta_M2M100_1_2B_AsyncResponse;
+interface Ai_Cf_Meta_M2M100_1_2B_AsyncResponse {
+  /**
+   * The async request id that can be used to obtain the results.
+   */
+  request_id?: string;
+}
 declare abstract class Base_Ai_Cf_Meta_M2M100_1_2B {
   inputs: Ai_Cf_Meta_M2M100_1_2B_Input;
   postProcessedOutputs: Ai_Cf_Meta_M2M100_1_2B_Output;
@@ -4261,7 +4688,13 @@ type Ai_Cf_Baai_Bge_Small_En_V1_5_Output =
        */
       pooling?: "mean" | "cls";
     }
-  | AsyncResponse;
+  | Ai_Cf_Baai_Bge_Small_En_V1_5_AsyncResponse;
+interface Ai_Cf_Baai_Bge_Small_En_V1_5_AsyncResponse {
+  /**
+   * The async request id that can be used to obtain the results.
+   */
+  request_id?: string;
+}
 declare abstract class Base_Ai_Cf_Baai_Bge_Small_En_V1_5 {
   inputs: Ai_Cf_Baai_Bge_Small_En_V1_5_Input;
   postProcessedOutputs: Ai_Cf_Baai_Bge_Small_En_V1_5_Output;
@@ -4298,7 +4731,13 @@ type Ai_Cf_Baai_Bge_Large_En_V1_5_Output =
        */
       pooling?: "mean" | "cls";
     }
-  | AsyncResponse;
+  | Ai_Cf_Baai_Bge_Large_En_V1_5_AsyncResponse;
+interface Ai_Cf_Baai_Bge_Large_En_V1_5_AsyncResponse {
+  /**
+   * The async request id that can be used to obtain the results.
+   */
+  request_id?: string;
+}
 declare abstract class Base_Ai_Cf_Baai_Bge_Large_En_V1_5 {
   inputs: Ai_Cf_Baai_Bge_Large_En_V1_5_Input;
   postProcessedOutputs: Ai_Cf_Baai_Bge_Large_En_V1_5_Output;
@@ -4489,15 +4928,18 @@ declare abstract class Base_Ai_Cf_Openai_Whisper_Large_V3_Turbo {
   postProcessedOutputs: Ai_Cf_Openai_Whisper_Large_V3_Turbo_Output;
 }
 type Ai_Cf_Baai_Bge_M3_Input =
-  | BGEM3InputQueryAndContexts
-  | BGEM3InputEmbedding
+  | Ai_Cf_Baai_Bge_M3_Input_QueryAnd_Contexts
+  | Ai_Cf_Baai_Bge_M3_Input_Embedding
   | {
       /**
        * Batch of the embeddings requests to run using async-queue
        */
-      requests: (BGEM3InputQueryAndContexts1 | BGEM3InputEmbedding1)[];
+      requests: (
+        | Ai_Cf_Baai_Bge_M3_Input_QueryAnd_Contexts_1
+        | Ai_Cf_Baai_Bge_M3_Input_Embedding_1
+      )[];
     };
-interface BGEM3InputQueryAndContexts {
+interface Ai_Cf_Baai_Bge_M3_Input_QueryAnd_Contexts {
   /**
    * A query you wish to perform against the provided contexts. If no query is provided the model with respond with embeddings for contexts
    */
@@ -4516,14 +4958,14 @@ interface BGEM3InputQueryAndContexts {
    */
   truncate_inputs?: boolean;
 }
-interface BGEM3InputEmbedding {
+interface Ai_Cf_Baai_Bge_M3_Input_Embedding {
   text: string | string[];
   /**
    * When provided with too long context should the model error out or truncate the context to fit?
    */
   truncate_inputs?: boolean;
 }
-interface BGEM3InputQueryAndContexts1 {
+interface Ai_Cf_Baai_Bge_M3_Input_QueryAnd_Contexts_1 {
   /**
    * A query you wish to perform against the provided contexts. If no query is provided the model with respond with embeddings for contexts
    */
@@ -4542,7 +4984,7 @@ interface BGEM3InputQueryAndContexts1 {
    */
   truncate_inputs?: boolean;
 }
-interface BGEM3InputEmbedding1 {
+interface Ai_Cf_Baai_Bge_M3_Input_Embedding_1 {
   text: string | string[];
   /**
    * When provided with too long context should the model error out or truncate the context to fit?
@@ -4550,11 +4992,11 @@ interface BGEM3InputEmbedding1 {
   truncate_inputs?: boolean;
 }
 type Ai_Cf_Baai_Bge_M3_Output =
-  | BGEM3OuputQuery
-  | BGEM3OutputEmbeddingForContexts
-  | BGEM3OuputEmbedding
-  | AsyncResponse;
-interface BGEM3OuputQuery {
+  | Ai_Cf_Baai_Bge_M3_Ouput_Query
+  | Ai_Cf_Baai_Bge_M3_Output_EmbeddingFor_Contexts
+  | Ai_Cf_Baai_Bge_M3_Ouput_Embedding
+  | Ai_Cf_Baai_Bge_M3_AsyncResponse;
+interface Ai_Cf_Baai_Bge_M3_Ouput_Query {
   response?: {
     /**
      * Index of the context in the request
@@ -4566,7 +5008,7 @@ interface BGEM3OuputQuery {
     score?: number;
   }[];
 }
-interface BGEM3OutputEmbeddingForContexts {
+interface Ai_Cf_Baai_Bge_M3_Output_EmbeddingFor_Contexts {
   response?: number[][];
   shape?: number[];
   /**
@@ -4574,7 +5016,7 @@ interface BGEM3OutputEmbeddingForContexts {
    */
   pooling?: "mean" | "cls";
 }
-interface BGEM3OuputEmbedding {
+interface Ai_Cf_Baai_Bge_M3_Ouput_Embedding {
   shape?: number[];
   /**
    * Embeddings of the requested text values
@@ -4585,6 +5027,12 @@ interface BGEM3OuputEmbedding {
    */
   pooling?: "mean" | "cls";
 }
+interface Ai_Cf_Baai_Bge_M3_AsyncResponse {
+  /**
+   * The async request id that can be used to obtain the results.
+   */
+  request_id?: string;
+}
 declare abstract class Base_Ai_Cf_Baai_Bge_M3 {
   inputs: Ai_Cf_Baai_Bge_M3_Input;
   postProcessedOutputs: Ai_Cf_Baai_Bge_M3_Output;
@@ -4609,8 +5057,10 @@ declare abstract class Base_Ai_Cf_Black_Forest_Labs_Flux_1_Schnell {
   inputs: Ai_Cf_Black_Forest_Labs_Flux_1_Schnell_Input;
   postProcessedOutputs: Ai_Cf_Black_Forest_Labs_Flux_1_Schnell_Output;
 }
-type Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct_Input = Prompt | Messages;
-interface Prompt {
+type Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct_Input =
+  | Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct_Prompt
+  | Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct_Messages;
+interface Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct_Prompt {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -4661,7 +5111,7 @@ interface Prompt {
    */
   lora?: string;
 }
-interface Messages {
+interface Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct_Messages {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -4859,10 +5309,10 @@ declare abstract class Base_Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct {
   postProcessedOutputs: Ai_Cf_Meta_Llama_3_2_11B_Vision_Instruct_Output;
 }
 type Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Input =
-  | Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Prompt
-  | Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Messages
-  | AsyncBatch;
-interface Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Prompt {
+  | Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Prompt
+  | Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Messages
+  | Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Async_Batch;
+interface Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Prompt {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -4871,7 +5321,7 @@ interface Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Prompt {
    * Name of the LoRA (Low-Rank Adaptation) model to fine-tune the base model.
    */
   lora?: string;
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_JSON_Mode;
   /**
    * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
@@ -4913,11 +5363,11 @@ interface Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Prompt {
    */
   presence_penalty?: number;
 }
-interface JSONMode {
+interface Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_JSON_Mode {
   type?: "json_object" | "json_schema";
   json_schema?: unknown;
 }
-interface Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Messages {
+interface Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Messages {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -5025,7 +5475,7 @@ interface Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Messages {
         };
       }
   )[];
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_JSON_Mode_1;
   /**
    * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
@@ -5067,7 +5517,11 @@ interface Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Messages {
    */
   presence_penalty?: number;
 }
-interface AsyncBatch {
+interface Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_JSON_Mode_1 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Async_Batch {
   requests?: {
     /**
      * User-supplied reference. This field will be present in the response as well it can be used to reference the request and response. It's NOT validated to be unique.
@@ -5109,9 +5563,13 @@ interface AsyncBatch {
      * Increases the likelihood of the model introducing new topics.
      */
     presence_penalty?: number;
-    response_format?: JSONMode;
+    response_format?: Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_JSON_Mode_2;
   }[];
 }
+interface Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_JSON_Mode_2 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
 type Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Output =
   | {
       /**
@@ -5150,7 +5608,13 @@ type Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Output =
       }[];
     }
   | string
-  | AsyncResponse;
+  | Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_AsyncResponse;
+interface Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_AsyncResponse {
+  /**
+   * The async request id that can be used to obtain the results.
+   */
+  request_id?: string;
+}
 declare abstract class Base_Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast {
   inputs: Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Input;
   postProcessedOutputs: Ai_Cf_Meta_Llama_3_3_70B_Instruct_Fp8_Fast_Output;
@@ -5257,9 +5721,9 @@ declare abstract class Base_Ai_Cf_Baai_Bge_Reranker_Base {
   postProcessedOutputs: Ai_Cf_Baai_Bge_Reranker_Base_Output;
 }
 type Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Input =
-  | Qwen2_5_Coder_32B_Instruct_Prompt
-  | Qwen2_5_Coder_32B_Instruct_Messages;
-interface Qwen2_5_Coder_32B_Instruct_Prompt {
+  | Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Prompt
+  | Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Messages;
+interface Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Prompt {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -5268,7 +5732,7 @@ interface Qwen2_5_Coder_32B_Instruct_Prompt {
    * Name of the LoRA (Low-Rank Adaptation) model to fine-tune the base model.
    */
   lora?: string;
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_JSON_Mode;
   /**
    * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
@@ -5310,7 +5774,11 @@ interface Qwen2_5_Coder_32B_Instruct_Prompt {
    */
   presence_penalty?: number;
 }
-interface Qwen2_5_Coder_32B_Instruct_Messages {
+interface Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_JSON_Mode {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Messages {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -5418,7 +5886,7 @@ interface Qwen2_5_Coder_32B_Instruct_Messages {
         };
       }
   )[];
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_JSON_Mode_1;
   /**
    * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
@@ -5460,6 +5928,10 @@ interface Qwen2_5_Coder_32B_Instruct_Messages {
    */
   presence_penalty?: number;
 }
+interface Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_JSON_Mode_1 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
 type Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Output = {
   /**
    * The generated text response from the model
@@ -5500,8 +5972,10 @@ declare abstract class Base_Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct {
   inputs: Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Input;
   postProcessedOutputs: Ai_Cf_Qwen_Qwen2_5_Coder_32B_Instruct_Output;
 }
-type Ai_Cf_Qwen_Qwq_32B_Input = Qwen_Qwq_32B_Prompt | Qwen_Qwq_32B_Messages;
-interface Qwen_Qwq_32B_Prompt {
+type Ai_Cf_Qwen_Qwq_32B_Input =
+  | Ai_Cf_Qwen_Qwq_32B_Prompt
+  | Ai_Cf_Qwen_Qwq_32B_Messages;
+interface Ai_Cf_Qwen_Qwq_32B_Prompt {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -5551,7 +6025,7 @@ interface Qwen_Qwq_32B_Prompt {
    */
   presence_penalty?: number;
 }
-interface Qwen_Qwq_32B_Messages {
+interface Ai_Cf_Qwen_Qwq_32B_Messages {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -5773,9 +6247,9 @@ declare abstract class Base_Ai_Cf_Qwen_Qwq_32B {
   postProcessedOutputs: Ai_Cf_Qwen_Qwq_32B_Output;
 }
 type Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct_Input =
-  | Mistral_Small_3_1_24B_Instruct_Prompt
-  | Mistral_Small_3_1_24B_Instruct_Messages;
-interface Mistral_Small_3_1_24B_Instruct_Prompt {
+  | Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct_Prompt
+  | Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct_Messages;
+interface Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct_Prompt {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -5825,7 +6299,7 @@ interface Mistral_Small_3_1_24B_Instruct_Prompt {
    */
   presence_penalty?: number;
 }
-interface Mistral_Small_3_1_24B_Instruct_Messages {
+interface Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct_Messages {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -6047,9 +6521,9 @@ declare abstract class Base_Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct {
   postProcessedOutputs: Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct_Output;
 }
 type Ai_Cf_Google_Gemma_3_12B_It_Input =
-  | Google_Gemma_3_12B_It_Prompt
-  | Google_Gemma_3_12B_It_Messages;
-interface Google_Gemma_3_12B_It_Prompt {
+  | Ai_Cf_Google_Gemma_3_12B_It_Prompt
+  | Ai_Cf_Google_Gemma_3_12B_It_Messages;
+interface Ai_Cf_Google_Gemma_3_12B_It_Prompt {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -6099,7 +6573,7 @@ interface Google_Gemma_3_12B_It_Prompt {
    */
   presence_penalty?: number;
 }
-interface Google_Gemma_3_12B_It_Messages {
+interface Ai_Cf_Google_Gemma_3_12B_It_Messages {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -6122,20 +6596,7 @@ interface Google_Gemma_3_12B_It_Messages {
              */
             url?: string;
           };
-        }[]
-      | {
-          /**
-           * Type of the content provided
-           */
-          type?: string;
-          text?: string;
-          image_url?: {
-            /**
-             * image uri with data (e.g. data:image/jpeg;base64,/9j/...). HTTP URL will not be accepted
-             */
-            url?: string;
-          };
-        };
+        }[];
   }[];
   functions?: {
     name: string;
@@ -6317,10 +6778,10 @@ declare abstract class Base_Ai_Cf_Google_Gemma_3_12B_It {
   postProcessedOutputs: Ai_Cf_Google_Gemma_3_12B_It_Output;
 }
 type Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Input =
-  | Ai_Cf_Meta_Llama_4_Prompt
-  | Ai_Cf_Meta_Llama_4_Messages
-  | Ai_Cf_Meta_Llama_4_Async_Batch;
-interface Ai_Cf_Meta_Llama_4_Prompt {
+  | Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Prompt
+  | Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Messages
+  | Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Async_Batch;
+interface Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Prompt {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -6329,7 +6790,7 @@ interface Ai_Cf_Meta_Llama_4_Prompt {
    * JSON schema that should be fulfilled for the response.
    */
   guided_json?: object;
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_JSON_Mode;
   /**
    * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
@@ -6371,7 +6832,11 @@ interface Ai_Cf_Meta_Llama_4_Prompt {
    */
   presence_penalty?: number;
 }
-interface Ai_Cf_Meta_Llama_4_Messages {
+interface Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_JSON_Mode {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Messages {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -6507,7 +6972,7 @@ interface Ai_Cf_Meta_Llama_4_Messages {
         };
       }
   )[];
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_JSON_Mode;
   /**
    * JSON schema that should be fufilled for the response.
    */
@@ -6553,13 +7018,13 @@ interface Ai_Cf_Meta_Llama_4_Messages {
    */
   presence_penalty?: number;
 }
-interface Ai_Cf_Meta_Llama_4_Async_Batch {
+interface Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Async_Batch {
   requests: (
-    | Ai_Cf_Meta_Llama_4_Prompt_Inner
-    | Ai_Cf_Meta_Llama_4_Messages_Inner
+    | Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Prompt_Inner
+    | Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Messages_Inner
   )[];
 }
-interface Ai_Cf_Meta_Llama_4_Prompt_Inner {
+interface Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Prompt_Inner {
   /**
    * The input text prompt for the model to generate a response.
    */
@@ -6568,7 +7033,7 @@ interface Ai_Cf_Meta_Llama_4_Prompt_Inner {
    * JSON schema that should be fulfilled for the response.
    */
   guided_json?: object;
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_JSON_Mode;
   /**
    * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
@@ -6610,7 +7075,7 @@ interface Ai_Cf_Meta_Llama_4_Prompt_Inner {
    */
   presence_penalty?: number;
 }
-interface Ai_Cf_Meta_Llama_4_Messages_Inner {
+interface Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Messages_Inner {
   /**
    * An array of message objects representing the conversation history.
    */
@@ -6746,7 +7211,7 @@ interface Ai_Cf_Meta_Llama_4_Messages_Inner {
         };
       }
   )[];
-  response_format?: JSONMode;
+  response_format?: Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_JSON_Mode;
   /**
    * JSON schema that should be fufilled for the response.
    */
@@ -6845,414 +7310,1813 @@ declare abstract class Base_Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct {
   inputs: Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Input;
   postProcessedOutputs: Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct_Output;
 }
-interface Ai_Cf_Deepgram_Nova_3_Input {
-  audio: {
-    body: object;
-    contentType: string;
-  };
-  /**
-   * Sets how the model will interpret strings submitted to the custom_topic param. When strict, the model will only return topics submitted using the custom_topic param. When extended, the model will return its own detected topics in addition to those submitted using the custom_topic param.
-   */
-  custom_topic_mode?: "extended" | "strict";
-  /**
-   * Custom topics you want the model to detect within your input audio or text if present Submit up to 100
-   */
-  custom_topic?: string;
-  /**
-   * Sets how the model will interpret intents submitted to the custom_intent param. When strict, the model will only return intents submitted using the custom_intent param. When extended, the model will return its own detected intents in addition those submitted using the custom_intents param
-   */
-  custom_intent_mode?: "extended" | "strict";
-  /**
-   * Custom intents you want the model to detect within your input audio if present
-   */
-  custom_intent?: string;
-  /**
-   * Identifies and extracts key entities from content in submitted audio
-   */
-  detect_entities?: boolean;
-  /**
-   * Identifies the dominant language spoken in submitted audio
-   */
-  detect_language?: boolean;
-  /**
-   * Recognize speaker changes. Each word in the transcript will be assigned a speaker number starting at 0
-   */
-  diarize?: boolean;
+type Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Input =
+  | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Prompt
+  | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Messages
+  | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Async_Batch;
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Prompt {
   /**
-   * Identify and extract key entities from content in submitted audio
+   * The input text prompt for the model to generate a response.
    */
-  dictation?: boolean;
+  prompt: string;
   /**
-   * Specify the expected encoding of your submitted audio
+   * Name of the LoRA (Low-Rank Adaptation) model to fine-tune the base model.
    */
-  encoding?:
-    | "linear16"
-    | "flac"
-    | "mulaw"
-    | "amr-nb"
-    | "amr-wb"
-    | "opus"
-    | "speex"
-    | "g729";
+  lora?: string;
+  response_format?: Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode;
   /**
-   * Arbitrary key-value pairs that are attached to the API response for usage in downstream processing
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
-  extra?: string;
+  raw?: boolean;
   /**
-   * Filler Words can help transcribe interruptions in your audio, like 'uh' and 'um'
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
    */
-  filler_words?: boolean;
+  stream?: boolean;
   /**
-   * Key term prompting can boost or suppress specialized terminology and brands.
+   * The maximum number of tokens to generate in the response.
    */
-  keyterm?: string;
+  max_tokens?: number;
   /**
-   * Keywords can boost or suppress specialized terminology and brands.
+   * Controls the randomness of the output; higher values produce more random results.
    */
-  keywords?: string;
+  temperature?: number;
   /**
-   * The BCP-47 language tag that hints at the primary spoken language. Depending on the Model and API endpoint you choose only certain languages are available.
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
    */
-  language?: string;
+  top_p?: number;
   /**
-   * Spoken measurements will be converted to their corresponding abbreviations.
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
    */
-  measurements?: boolean;
+  top_k?: number;
   /**
-   * Opts out requests from the Deepgram Model Improvement Program. Refer to our Docs for pricing impacts before setting this to true. https://dpgr.am/deepgram-mip.
+   * Random seed for reproducibility of the generation.
    */
-  mip_opt_out?: boolean;
+  seed?: number;
   /**
-   * Mode of operation for the model representing broad area of topic that will be talked about in the supplied audio
+   * Penalty for repeated tokens; higher values discourage repetition.
    */
-  mode?: "general" | "medical" | "finance";
+  repetition_penalty?: number;
   /**
-   * Transcribe each audio channel independently.
+   * Decreases the likelihood of the model repeating the same lines verbatim.
    */
-  multichannel?: boolean;
+  frequency_penalty?: number;
   /**
-   * Numerals converts numbers from written format to numerical format.
+   * Increases the likelihood of the model introducing new topics.
    */
-  numerals?: boolean;
+  presence_penalty?: number;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Messages {
   /**
-   * Splits audio into paragraphs to improve transcript readability.
+   * An array of message objects representing the conversation history.
    */
-  paragraphs?: boolean;
+  messages: {
+    /**
+     * The role of the message sender (e.g., 'user', 'assistant', 'system', 'tool').
+     */
+    role: string;
+    /**
+     * The content of the message as a string.
+     */
+    content: string;
+  }[];
+  functions?: {
+    name: string;
+    code: string;
+  }[];
   /**
-   * Profanity Filter looks for recognized profanity and converts it to the nearest recognized non-profane word or removes it from the transcript completely.
+   * A list of tools available for the assistant to use.
+   */
+  tools?: (
+    | {
+        /**
+         * The name of the tool. More descriptive the better.
+         */
+        name: string;
+        /**
+         * A brief description of what the tool does.
+         */
+        description: string;
+        /**
+         * Schema defining the parameters accepted by the tool.
+         */
+        parameters: {
+          /**
+           * The type of the parameters object (usually 'object').
+           */
+          type: string;
+          /**
+           * List of required parameter names.
+           */
+          required?: string[];
+          /**
+           * Definitions of each parameter.
+           */
+          properties: {
+            [k: string]: {
+              /**
+               * The data type of the parameter.
+               */
+              type: string;
+              /**
+               * A description of the expected parameter.
+               */
+              description: string;
+            };
+          };
+        };
+      }
+    | {
+        /**
+         * Specifies the type of tool (e.g., 'function').
+         */
+        type: string;
+        /**
+         * Details of the function tool.
+         */
+        function: {
+          /**
+           * The name of the function.
+           */
+          name: string;
+          /**
+           * A brief description of what the function does.
+           */
+          description: string;
+          /**
+           * Schema defining the parameters accepted by the function.
+           */
+          parameters: {
+            /**
+             * The type of the parameters object (usually 'object').
+             */
+            type: string;
+            /**
+             * List of required parameter names.
+             */
+            required?: string[];
+            /**
+             * Definitions of each parameter.
+             */
+            properties: {
+              [k: string]: {
+                /**
+                 * The data type of the parameter.
+                 */
+                type: string;
+                /**
+                 * A description of the expected parameter.
+                 */
+                description: string;
+              };
+            };
+          };
+        };
+      }
+  )[];
+  response_format?: Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode_1;
+  /**
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
+   */
+  raw?: boolean;
+  /**
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
+   */
+  stream?: boolean;
+  /**
+   * The maximum number of tokens to generate in the response.
+   */
+  max_tokens?: number;
+  /**
+   * Controls the randomness of the output; higher values produce more random results.
+   */
+  temperature?: number;
+  /**
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
+   */
+  top_p?: number;
+  /**
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
+   */
+  top_k?: number;
+  /**
+   * Random seed for reproducibility of the generation.
+   */
+  seed?: number;
+  /**
+   * Penalty for repeated tokens; higher values discourage repetition.
+   */
+  repetition_penalty?: number;
+  /**
+   * Decreases the likelihood of the model repeating the same lines verbatim.
+   */
+  frequency_penalty?: number;
+  /**
+   * Increases the likelihood of the model introducing new topics.
+   */
+  presence_penalty?: number;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode_1 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Async_Batch {
+  requests: (
+    | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Prompt_1
+    | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Messages_1
+  )[];
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Prompt_1 {
+  /**
+   * The input text prompt for the model to generate a response.
+   */
+  prompt: string;
+  /**
+   * Name of the LoRA (Low-Rank Adaptation) model to fine-tune the base model.
+   */
+  lora?: string;
+  response_format?: Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode_2;
+  /**
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
+   */
+  raw?: boolean;
+  /**
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
+   */
+  stream?: boolean;
+  /**
+   * The maximum number of tokens to generate in the response.
+   */
+  max_tokens?: number;
+  /**
+   * Controls the randomness of the output; higher values produce more random results.
+   */
+  temperature?: number;
+  /**
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
+   */
+  top_p?: number;
+  /**
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
+   */
+  top_k?: number;
+  /**
+   * Random seed for reproducibility of the generation.
+   */
+  seed?: number;
+  /**
+   * Penalty for repeated tokens; higher values discourage repetition.
+   */
+  repetition_penalty?: number;
+  /**
+   * Decreases the likelihood of the model repeating the same lines verbatim.
+   */
+  frequency_penalty?: number;
+  /**
+   * Increases the likelihood of the model introducing new topics.
+   */
+  presence_penalty?: number;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode_2 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Messages_1 {
+  /**
+   * An array of message objects representing the conversation history.
+   */
+  messages: {
+    /**
+     * The role of the message sender (e.g., 'user', 'assistant', 'system', 'tool').
+     */
+    role: string;
+    /**
+     * The content of the message as a string.
+     */
+    content: string;
+  }[];
+  functions?: {
+    name: string;
+    code: string;
+  }[];
+  /**
+   * A list of tools available for the assistant to use.
+   */
+  tools?: (
+    | {
+        /**
+         * The name of the tool. More descriptive the better.
+         */
+        name: string;
+        /**
+         * A brief description of what the tool does.
+         */
+        description: string;
+        /**
+         * Schema defining the parameters accepted by the tool.
+         */
+        parameters: {
+          /**
+           * The type of the parameters object (usually 'object').
+           */
+          type: string;
+          /**
+           * List of required parameter names.
+           */
+          required?: string[];
+          /**
+           * Definitions of each parameter.
+           */
+          properties: {
+            [k: string]: {
+              /**
+               * The data type of the parameter.
+               */
+              type: string;
+              /**
+               * A description of the expected parameter.
+               */
+              description: string;
+            };
+          };
+        };
+      }
+    | {
+        /**
+         * Specifies the type of tool (e.g., 'function').
+         */
+        type: string;
+        /**
+         * Details of the function tool.
+         */
+        function: {
+          /**
+           * The name of the function.
+           */
+          name: string;
+          /**
+           * A brief description of what the function does.
+           */
+          description: string;
+          /**
+           * Schema defining the parameters accepted by the function.
+           */
+          parameters: {
+            /**
+             * The type of the parameters object (usually 'object').
+             */
+            type: string;
+            /**
+             * List of required parameter names.
+             */
+            required?: string[];
+            /**
+             * Definitions of each parameter.
+             */
+            properties: {
+              [k: string]: {
+                /**
+                 * The data type of the parameter.
+                 */
+                type: string;
+                /**
+                 * A description of the expected parameter.
+                 */
+                description: string;
+              };
+            };
+          };
+        };
+      }
+  )[];
+  response_format?: Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode_3;
+  /**
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
+   */
+  raw?: boolean;
+  /**
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
+   */
+  stream?: boolean;
+  /**
+   * The maximum number of tokens to generate in the response.
+   */
+  max_tokens?: number;
+  /**
+   * Controls the randomness of the output; higher values produce more random results.
+   */
+  temperature?: number;
+  /**
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
+   */
+  top_p?: number;
+  /**
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
+   */
+  top_k?: number;
+  /**
+   * Random seed for reproducibility of the generation.
+   */
+  seed?: number;
+  /**
+   * Penalty for repeated tokens; higher values discourage repetition.
+   */
+  repetition_penalty?: number;
+  /**
+   * Decreases the likelihood of the model repeating the same lines verbatim.
+   */
+  frequency_penalty?: number;
+  /**
+   * Increases the likelihood of the model introducing new topics.
+   */
+  presence_penalty?: number;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_JSON_Mode_3 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+type Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Output =
+  | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Chat_Completion_Response
+  | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Text_Completion_Response
+  | string
+  | Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_AsyncResponse;
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Chat_Completion_Response {
+  /**
+   * Unique identifier for the completion
+   */
+  id?: string;
+  /**
+   * Object type identifier
+   */
+  object?: "chat.completion";
+  /**
+   * Unix timestamp of when the completion was created
+   */
+  created?: number;
+  /**
+   * Model used for the completion
+   */
+  model?: string;
+  /**
+   * List of completion choices
+   */
+  choices?: {
+    /**
+     * Index of the choice in the list
+     */
+    index?: number;
+    /**
+     * The message generated by the model
+     */
+    message?: {
+      /**
+       * Role of the message author
+       */
+      role: string;
+      /**
+       * The content of the message
+       */
+      content: string;
+      /**
+       * Internal reasoning content (if available)
+       */
+      reasoning_content?: string;
+      /**
+       * Tool calls made by the assistant
+       */
+      tool_calls?: {
+        /**
+         * Unique identifier for the tool call
+         */
+        id: string;
+        /**
+         * Type of tool call
+         */
+        type: "function";
+        function: {
+          /**
+           * Name of the function to call
+           */
+          name: string;
+          /**
+           * JSON string of arguments for the function
+           */
+          arguments: string;
+        };
+      }[];
+    };
+    /**
+     * Reason why the model stopped generating
+     */
+    finish_reason?: string;
+    /**
+     * Stop reason (may be null)
+     */
+    stop_reason?: string | null;
+    /**
+     * Log probabilities (if requested)
+     */
+    logprobs?: {} | null;
+  }[];
+  /**
+   * Usage statistics for the inference request
+   */
+  usage?: {
+    /**
+     * Total number of tokens in input
+     */
+    prompt_tokens?: number;
+    /**
+     * Total number of tokens in output
+     */
+    completion_tokens?: number;
+    /**
+     * Total number of input and output tokens
+     */
+    total_tokens?: number;
+  };
+  /**
+   * Log probabilities for the prompt (if requested)
+   */
+  prompt_logprobs?: {} | null;
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Text_Completion_Response {
+  /**
+   * Unique identifier for the completion
+   */
+  id?: string;
+  /**
+   * Object type identifier
+   */
+  object?: "text_completion";
+  /**
+   * Unix timestamp of when the completion was created
+   */
+  created?: number;
+  /**
+   * Model used for the completion
+   */
+  model?: string;
+  /**
+   * List of completion choices
+   */
+  choices?: {
+    /**
+     * Index of the choice in the list
+     */
+    index: number;
+    /**
+     * The generated text completion
+     */
+    text: string;
+    /**
+     * Reason why the model stopped generating
+     */
+    finish_reason: string;
+    /**
+     * Stop reason (may be null)
+     */
+    stop_reason?: string | null;
+    /**
+     * Log probabilities (if requested)
+     */
+    logprobs?: {} | null;
+    /**
+     * Log probabilities for the prompt (if requested)
+     */
+    prompt_logprobs?: {} | null;
+  }[];
+  /**
+   * Usage statistics for the inference request
+   */
+  usage?: {
+    /**
+     * Total number of tokens in input
+     */
+    prompt_tokens?: number;
+    /**
+     * Total number of tokens in output
+     */
+    completion_tokens?: number;
+    /**
+     * Total number of input and output tokens
+     */
+    total_tokens?: number;
+  };
+}
+interface Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_AsyncResponse {
+  /**
+   * The async request id that can be used to obtain the results.
+   */
+  request_id?: string;
+}
+declare abstract class Base_Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8 {
+  inputs: Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Input;
+  postProcessedOutputs: Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8_Output;
+}
+interface Ai_Cf_Deepgram_Nova_3_Input {
+  audio: {
+    body: object;
+    contentType: string;
+  };
+  /**
+   * Sets how the model will interpret strings submitted to the custom_topic param. When strict, the model will only return topics submitted using the custom_topic param. When extended, the model will return its own detected topics in addition to those submitted using the custom_topic param.
+   */
+  custom_topic_mode?: "extended" | "strict";
+  /**
+   * Custom topics you want the model to detect within your input audio or text if present Submit up to 100
+   */
+  custom_topic?: string;
+  /**
+   * Sets how the model will interpret intents submitted to the custom_intent param. When strict, the model will only return intents submitted using the custom_intent param. When extended, the model will return its own detected intents in addition those submitted using the custom_intents param
+   */
+  custom_intent_mode?: "extended" | "strict";
+  /**
+   * Custom intents you want the model to detect within your input audio if present
+   */
+  custom_intent?: string;
+  /**
+   * Identifies and extracts key entities from content in submitted audio
+   */
+  detect_entities?: boolean;
+  /**
+   * Identifies the dominant language spoken in submitted audio
+   */
+  detect_language?: boolean;
+  /**
+   * Recognize speaker changes. Each word in the transcript will be assigned a speaker number starting at 0
+   */
+  diarize?: boolean;
+  /**
+   * Identify and extract key entities from content in submitted audio
+   */
+  dictation?: boolean;
+  /**
+   * Specify the expected encoding of your submitted audio
+   */
+  encoding?:
+    | "linear16"
+    | "flac"
+    | "mulaw"
+    | "amr-nb"
+    | "amr-wb"
+    | "opus"
+    | "speex"
+    | "g729";
+  /**
+   * Arbitrary key-value pairs that are attached to the API response for usage in downstream processing
+   */
+  extra?: string;
+  /**
+   * Filler Words can help transcribe interruptions in your audio, like 'uh' and 'um'
+   */
+  filler_words?: boolean;
+  /**
+   * Key term prompting can boost or suppress specialized terminology and brands.
+   */
+  keyterm?: string;
+  /**
+   * Keywords can boost or suppress specialized terminology and brands.
+   */
+  keywords?: string;
+  /**
+   * The BCP-47 language tag that hints at the primary spoken language. Depending on the Model and API endpoint you choose only certain languages are available.
+   */
+  language?: string;
+  /**
+   * Spoken measurements will be converted to their corresponding abbreviations.
+   */
+  measurements?: boolean;
+  /**
+   * Opts out requests from the Deepgram Model Improvement Program. Refer to our Docs for pricing impacts before setting this to true. https://dpgr.am/deepgram-mip.
+   */
+  mip_opt_out?: boolean;
+  /**
+   * Mode of operation for the model representing broad area of topic that will be talked about in the supplied audio
+   */
+  mode?: "general" | "medical" | "finance";
+  /**
+   * Transcribe each audio channel independently.
+   */
+  multichannel?: boolean;
+  /**
+   * Numerals converts numbers from written format to numerical format.
+   */
+  numerals?: boolean;
+  /**
+   * Splits audio into paragraphs to improve transcript readability.
+   */
+  paragraphs?: boolean;
+  /**
+   * Profanity Filter looks for recognized profanity and converts it to the nearest recognized non-profane word or removes it from the transcript completely.
    */
   profanity_filter?: boolean;
   /**
-   * Add punctuation and capitalization to the transcript.
+   * Add punctuation and capitalization to the transcript.
+   */
+  punctuate?: boolean;
+  /**
+   * Redaction removes sensitive information from your transcripts.
+   */
+  redact?: string;
+  /**
+   * Search for terms or phrases in submitted audio and replaces them.
+   */
+  replace?: string;
+  /**
+   * Search for terms or phrases in submitted audio.
+   */
+  search?: string;
+  /**
+   * Recognizes the sentiment throughout a transcript or text.
+   */
+  sentiment?: boolean;
+  /**
+   * Apply formatting to transcript output. When set to true, additional formatting will be applied to transcripts to improve readability.
+   */
+  smart_format?: boolean;
+  /**
+   * Detect topics throughout a transcript or text.
+   */
+  topics?: boolean;
+  /**
+   * Segments speech into meaningful semantic units.
+   */
+  utterances?: boolean;
+  /**
+   * Seconds to wait before detecting a pause between words in submitted audio.
+   */
+  utt_split?: number;
+  /**
+   * The number of channels in the submitted audio
+   */
+  channels?: number;
+  /**
+   * Specifies whether the streaming endpoint should provide ongoing transcription updates as more audio is received. When set to true, the endpoint sends continuous updates, meaning transcription results may evolve over time. Note: Supported only for webosockets.
+   */
+  interim_results?: boolean;
+  /**
+   * Indicates how long model will wait to detect whether a speaker has finished speaking or pauses for a significant period of time. When set to a value, the streaming endpoint immediately finalizes the transcription for the processed time range and returns the transcript with a speech_final parameter set to true. Can also be set to false to disable endpointing
+   */
+  endpointing?: string;
+  /**
+   * Indicates that speech has started. You'll begin receiving Speech Started messages upon speech starting. Note: Supported only for webosockets.
+   */
+  vad_events?: boolean;
+  /**
+   * Indicates how long model will wait to send an UtteranceEnd message after a word has been transcribed. Use with interim_results. Note: Supported only for webosockets.
+   */
+  utterance_end_ms?: boolean;
+}
+interface Ai_Cf_Deepgram_Nova_3_Output {
+  results?: {
+    channels?: {
+      alternatives?: {
+        confidence?: number;
+        transcript?: string;
+        words?: {
+          confidence?: number;
+          end?: number;
+          start?: number;
+          word?: string;
+        }[];
+      }[];
+    }[];
+    summary?: {
+      result?: string;
+      short?: string;
+    };
+    sentiments?: {
+      segments?: {
+        text?: string;
+        start_word?: number;
+        end_word?: number;
+        sentiment?: string;
+        sentiment_score?: number;
+      }[];
+      average?: {
+        sentiment?: string;
+        sentiment_score?: number;
+      };
+    };
+  };
+}
+declare abstract class Base_Ai_Cf_Deepgram_Nova_3 {
+  inputs: Ai_Cf_Deepgram_Nova_3_Input;
+  postProcessedOutputs: Ai_Cf_Deepgram_Nova_3_Output;
+}
+interface Ai_Cf_Qwen_Qwen3_Embedding_0_6B_Input {
+  queries?: string | string[];
+  /**
+   * Optional instruction for the task
+   */
+  instruction?: string;
+  documents?: string | string[];
+  text?: string | string[];
+}
+interface Ai_Cf_Qwen_Qwen3_Embedding_0_6B_Output {
+  data?: number[][];
+  shape?: number[];
+}
+declare abstract class Base_Ai_Cf_Qwen_Qwen3_Embedding_0_6B {
+  inputs: Ai_Cf_Qwen_Qwen3_Embedding_0_6B_Input;
+  postProcessedOutputs: Ai_Cf_Qwen_Qwen3_Embedding_0_6B_Output;
+}
+type Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Input =
+  | {
+      /**
+       * readable stream with audio data and content-type specified for that data
+       */
+      audio: {
+        body: object;
+        contentType: string;
+      };
+      /**
+       * type of data PCM data that's sent to the inference server as raw array
+       */
+      dtype?: "uint8" | "float32" | "float64";
+    }
+  | {
+      /**
+       * base64 encoded audio data
+       */
+      audio: string;
+      /**
+       * type of data PCM data that's sent to the inference server as raw array
+       */
+      dtype?: "uint8" | "float32" | "float64";
+    };
+interface Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Output {
+  /**
+   * if true, end-of-turn was detected
+   */
+  is_complete?: boolean;
+  /**
+   * probability of the end-of-turn detection
+   */
+  probability?: number;
+}
+declare abstract class Base_Ai_Cf_Pipecat_Ai_Smart_Turn_V2 {
+  inputs: Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Input;
+  postProcessedOutputs: Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Output;
+}
+declare abstract class Base_Ai_Cf_Openai_Gpt_Oss_120B {
+  inputs: ResponsesInput;
+  postProcessedOutputs: ResponsesOutput;
+}
+declare abstract class Base_Ai_Cf_Openai_Gpt_Oss_20B {
+  inputs: ResponsesInput;
+  postProcessedOutputs: ResponsesOutput;
+}
+interface Ai_Cf_Leonardo_Phoenix_1_0_Input {
+  /**
+   * A text description of the image you want to generate.
+   */
+  prompt: string;
+  /**
+   * Controls how closely the generated image should adhere to the prompt; higher values make the image more aligned with the prompt
+   */
+  guidance?: number;
+  /**
+   * Random seed for reproducibility of the image generation
+   */
+  seed?: number;
+  /**
+   * The height of the generated image in pixels
+   */
+  height?: number;
+  /**
+   * The width of the generated image in pixels
+   */
+  width?: number;
+  /**
+   * The number of diffusion steps; higher values can improve quality but take longer
+   */
+  num_steps?: number;
+  /**
+   * Specify what to exclude from the generated images
+   */
+  negative_prompt?: string;
+}
+/**
+ * The generated image in JPEG format
+ */
+type Ai_Cf_Leonardo_Phoenix_1_0_Output = string;
+declare abstract class Base_Ai_Cf_Leonardo_Phoenix_1_0 {
+  inputs: Ai_Cf_Leonardo_Phoenix_1_0_Input;
+  postProcessedOutputs: Ai_Cf_Leonardo_Phoenix_1_0_Output;
+}
+interface Ai_Cf_Leonardo_Lucid_Origin_Input {
+  /**
+   * A text description of the image you want to generate.
+   */
+  prompt: string;
+  /**
+   * Controls how closely the generated image should adhere to the prompt; higher values make the image more aligned with the prompt
+   */
+  guidance?: number;
+  /**
+   * Random seed for reproducibility of the image generation
+   */
+  seed?: number;
+  /**
+   * The height of the generated image in pixels
+   */
+  height?: number;
+  /**
+   * The width of the generated image in pixels
+   */
+  width?: number;
+  /**
+   * The number of diffusion steps; higher values can improve quality but take longer
+   */
+  num_steps?: number;
+  /**
+   * The number of diffusion steps; higher values can improve quality but take longer
+   */
+  steps?: number;
+}
+interface Ai_Cf_Leonardo_Lucid_Origin_Output {
+  /**
+   * The generated image in Base64 format.
+   */
+  image?: string;
+}
+declare abstract class Base_Ai_Cf_Leonardo_Lucid_Origin {
+  inputs: Ai_Cf_Leonardo_Lucid_Origin_Input;
+  postProcessedOutputs: Ai_Cf_Leonardo_Lucid_Origin_Output;
+}
+interface Ai_Cf_Deepgram_Aura_1_Input {
+  /**
+   * Speaker used to produce the audio.
+   */
+  speaker?:
+    | "angus"
+    | "asteria"
+    | "arcas"
+    | "orion"
+    | "orpheus"
+    | "athena"
+    | "luna"
+    | "zeus"
+    | "perseus"
+    | "helios"
+    | "hera"
+    | "stella";
+  /**
+   * Encoding of the output audio.
+   */
+  encoding?: "linear16" | "flac" | "mulaw" | "alaw" | "mp3" | "opus" | "aac";
+  /**
+   * Container specifies the file format wrapper for the output audio. The available options depend on the encoding type..
+   */
+  container?: "none" | "wav" | "ogg";
+  /**
+   * The text content to be converted to speech
+   */
+  text: string;
+  /**
+   * Sample Rate specifies the sample rate for the output audio. Based on the encoding, different sample rates are supported. For some encodings, the sample rate is not configurable
+   */
+  sample_rate?: number;
+  /**
+   * The bitrate of the audio in bits per second. Choose from predefined ranges or specific values based on the encoding type.
+   */
+  bit_rate?: number;
+}
+/**
+ * The generated audio in MP3 format
+ */
+type Ai_Cf_Deepgram_Aura_1_Output = string;
+declare abstract class Base_Ai_Cf_Deepgram_Aura_1 {
+  inputs: Ai_Cf_Deepgram_Aura_1_Input;
+  postProcessedOutputs: Ai_Cf_Deepgram_Aura_1_Output;
+}
+interface Ai_Cf_Ai4Bharat_Indictrans2_En_Indic_1B_Input {
+  /**
+   * Input text to translate. Can be a single string or a list of strings.
+   */
+  text: string | string[];
+  /**
+   * Target langauge to translate to
+   */
+  target_language:
+    | "asm_Beng"
+    | "awa_Deva"
+    | "ben_Beng"
+    | "bho_Deva"
+    | "brx_Deva"
+    | "doi_Deva"
+    | "eng_Latn"
+    | "gom_Deva"
+    | "gon_Deva"
+    | "guj_Gujr"
+    | "hin_Deva"
+    | "hne_Deva"
+    | "kan_Knda"
+    | "kas_Arab"
+    | "kas_Deva"
+    | "kha_Latn"
+    | "lus_Latn"
+    | "mag_Deva"
+    | "mai_Deva"
+    | "mal_Mlym"
+    | "mar_Deva"
+    | "mni_Beng"
+    | "mni_Mtei"
+    | "npi_Deva"
+    | "ory_Orya"
+    | "pan_Guru"
+    | "san_Deva"
+    | "sat_Olck"
+    | "snd_Arab"
+    | "snd_Deva"
+    | "tam_Taml"
+    | "tel_Telu"
+    | "urd_Arab"
+    | "unr_Deva";
+}
+interface Ai_Cf_Ai4Bharat_Indictrans2_En_Indic_1B_Output {
+  /**
+   * Translated texts
+   */
+  translations: string[];
+}
+declare abstract class Base_Ai_Cf_Ai4Bharat_Indictrans2_En_Indic_1B {
+  inputs: Ai_Cf_Ai4Bharat_Indictrans2_En_Indic_1B_Input;
+  postProcessedOutputs: Ai_Cf_Ai4Bharat_Indictrans2_En_Indic_1B_Output;
+}
+type Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Input =
+  | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Prompt
+  | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Messages
+  | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Async_Batch;
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Prompt {
+  /**
+   * The input text prompt for the model to generate a response.
+   */
+  prompt: string;
+  /**
+   * Name of the LoRA (Low-Rank Adaptation) model to fine-tune the base model.
+   */
+  lora?: string;
+  response_format?: Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode;
+  /**
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
+   */
+  raw?: boolean;
+  /**
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
+   */
+  stream?: boolean;
+  /**
+   * The maximum number of tokens to generate in the response.
+   */
+  max_tokens?: number;
+  /**
+   * Controls the randomness of the output; higher values produce more random results.
+   */
+  temperature?: number;
+  /**
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
+   */
+  top_p?: number;
+  /**
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
+   */
+  top_k?: number;
+  /**
+   * Random seed for reproducibility of the generation.
+   */
+  seed?: number;
+  /**
+   * Penalty for repeated tokens; higher values discourage repetition.
+   */
+  repetition_penalty?: number;
+  /**
+   * Decreases the likelihood of the model repeating the same lines verbatim.
+   */
+  frequency_penalty?: number;
+  /**
+   * Increases the likelihood of the model introducing new topics.
+   */
+  presence_penalty?: number;
+}
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Messages {
+  /**
+   * An array of message objects representing the conversation history.
+   */
+  messages: {
+    /**
+     * The role of the message sender (e.g., 'user', 'assistant', 'system', 'tool').
+     */
+    role: string;
+    /**
+     * The content of the message as a string.
+     */
+    content: string;
+  }[];
+  functions?: {
+    name: string;
+    code: string;
+  }[];
+  /**
+   * A list of tools available for the assistant to use.
+   */
+  tools?: (
+    | {
+        /**
+         * The name of the tool. More descriptive the better.
+         */
+        name: string;
+        /**
+         * A brief description of what the tool does.
+         */
+        description: string;
+        /**
+         * Schema defining the parameters accepted by the tool.
+         */
+        parameters: {
+          /**
+           * The type of the parameters object (usually 'object').
+           */
+          type: string;
+          /**
+           * List of required parameter names.
+           */
+          required?: string[];
+          /**
+           * Definitions of each parameter.
+           */
+          properties: {
+            [k: string]: {
+              /**
+               * The data type of the parameter.
+               */
+              type: string;
+              /**
+               * A description of the expected parameter.
+               */
+              description: string;
+            };
+          };
+        };
+      }
+    | {
+        /**
+         * Specifies the type of tool (e.g., 'function').
+         */
+        type: string;
+        /**
+         * Details of the function tool.
+         */
+        function: {
+          /**
+           * The name of the function.
+           */
+          name: string;
+          /**
+           * A brief description of what the function does.
+           */
+          description: string;
+          /**
+           * Schema defining the parameters accepted by the function.
+           */
+          parameters: {
+            /**
+             * The type of the parameters object (usually 'object').
+             */
+            type: string;
+            /**
+             * List of required parameter names.
+             */
+            required?: string[];
+            /**
+             * Definitions of each parameter.
+             */
+            properties: {
+              [k: string]: {
+                /**
+                 * The data type of the parameter.
+                 */
+                type: string;
+                /**
+                 * A description of the expected parameter.
+                 */
+                description: string;
+              };
+            };
+          };
+        };
+      }
+  )[];
+  response_format?: Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode_1;
+  /**
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
+   */
+  raw?: boolean;
+  /**
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
+   */
+  stream?: boolean;
+  /**
+   * The maximum number of tokens to generate in the response.
+   */
+  max_tokens?: number;
+  /**
+   * Controls the randomness of the output; higher values produce more random results.
+   */
+  temperature?: number;
+  /**
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
    */
-  punctuate?: boolean;
+  top_p?: number;
   /**
-   * Redaction removes sensitive information from your transcripts.
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
    */
-  redact?: string;
+  top_k?: number;
   /**
-   * Search for terms or phrases in submitted audio and replaces them.
+   * Random seed for reproducibility of the generation.
    */
-  replace?: string;
+  seed?: number;
   /**
-   * Search for terms or phrases in submitted audio.
+   * Penalty for repeated tokens; higher values discourage repetition.
    */
-  search?: string;
+  repetition_penalty?: number;
   /**
-   * Recognizes the sentiment throughout a transcript or text.
+   * Decreases the likelihood of the model repeating the same lines verbatim.
    */
-  sentiment?: boolean;
+  frequency_penalty?: number;
   /**
-   * Apply formatting to transcript output. When set to true, additional formatting will be applied to transcripts to improve readability.
+   * Increases the likelihood of the model introducing new topics.
    */
-  smart_format?: boolean;
+  presence_penalty?: number;
+}
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode_1 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Async_Batch {
+  requests: (
+    | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Prompt_1
+    | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Messages_1
+  )[];
+}
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Prompt_1 {
   /**
-   * Detect topics throughout a transcript or text.
+   * The input text prompt for the model to generate a response.
    */
-  topics?: boolean;
+  prompt: string;
   /**
-   * Segments speech into meaningful semantic units.
+   * Name of the LoRA (Low-Rank Adaptation) model to fine-tune the base model.
    */
-  utterances?: boolean;
+  lora?: string;
+  response_format?: Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode_2;
   /**
-   * Seconds to wait before detecting a pause between words in submitted audio.
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
    */
-  utt_split?: number;
+  raw?: boolean;
   /**
-   * The number of channels in the submitted audio
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
    */
-  channels?: number;
+  stream?: boolean;
   /**
-   * Specifies whether the streaming endpoint should provide ongoing transcription updates as more audio is received. When set to true, the endpoint sends continuous updates, meaning transcription results may evolve over time. Note: Supported only for webosockets.
+   * The maximum number of tokens to generate in the response.
    */
-  interim_results?: boolean;
+  max_tokens?: number;
   /**
-   * Indicates how long model will wait to detect whether a speaker has finished speaking or pauses for a significant period of time. When set to a value, the streaming endpoint immediately finalizes the transcription for the processed time range and returns the transcript with a speech_final parameter set to true. Can also be set to false to disable endpointing
+   * Controls the randomness of the output; higher values produce more random results.
    */
-  endpointing?: string;
+  temperature?: number;
   /**
-   * Indicates that speech has started. You'll begin receiving Speech Started messages upon speech starting. Note: Supported only for webosockets.
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
    */
-  vad_events?: boolean;
+  top_p?: number;
   /**
-   * Indicates how long model will wait to send an UtteranceEnd message after a word has been transcribed. Use with interim_results. Note: Supported only for webosockets.
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
    */
-  utterance_end_ms?: boolean;
+  top_k?: number;
+  /**
+   * Random seed for reproducibility of the generation.
+   */
+  seed?: number;
+  /**
+   * Penalty for repeated tokens; higher values discourage repetition.
+   */
+  repetition_penalty?: number;
+  /**
+   * Decreases the likelihood of the model repeating the same lines verbatim.
+   */
+  frequency_penalty?: number;
+  /**
+   * Increases the likelihood of the model introducing new topics.
+   */
+  presence_penalty?: number;
 }
-interface Ai_Cf_Deepgram_Nova_3_Output {
-  results?: {
-    channels?: {
-      alternatives?: {
-        confidence?: number;
-        transcript?: string;
-        words?: {
-          confidence?: number;
-          end?: number;
-          start?: number;
-          word?: string;
-        }[];
-      }[];
-    }[];
-    summary?: {
-      result?: string;
-      short?: string;
-    };
-    sentiments?: {
-      segments?: {
-        text?: string;
-        start_word?: number;
-        end_word?: number;
-        sentiment?: string;
-        sentiment_score?: number;
-      }[];
-      average?: {
-        sentiment?: string;
-        sentiment_score?: number;
-      };
-    };
-  };
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode_2 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
+}
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Messages_1 {
+  /**
+   * An array of message objects representing the conversation history.
+   */
+  messages: {
+    /**
+     * The role of the message sender (e.g., 'user', 'assistant', 'system', 'tool').
+     */
+    role: string;
+    /**
+     * The content of the message as a string.
+     */
+    content: string;
+  }[];
+  functions?: {
+    name: string;
+    code: string;
+  }[];
+  /**
+   * A list of tools available for the assistant to use.
+   */
+  tools?: (
+    | {
+        /**
+         * The name of the tool. More descriptive the better.
+         */
+        name: string;
+        /**
+         * A brief description of what the tool does.
+         */
+        description: string;
+        /**
+         * Schema defining the parameters accepted by the tool.
+         */
+        parameters: {
+          /**
+           * The type of the parameters object (usually 'object').
+           */
+          type: string;
+          /**
+           * List of required parameter names.
+           */
+          required?: string[];
+          /**
+           * Definitions of each parameter.
+           */
+          properties: {
+            [k: string]: {
+              /**
+               * The data type of the parameter.
+               */
+              type: string;
+              /**
+               * A description of the expected parameter.
+               */
+              description: string;
+            };
+          };
+        };
+      }
+    | {
+        /**
+         * Specifies the type of tool (e.g., 'function').
+         */
+        type: string;
+        /**
+         * Details of the function tool.
+         */
+        function: {
+          /**
+           * The name of the function.
+           */
+          name: string;
+          /**
+           * A brief description of what the function does.
+           */
+          description: string;
+          /**
+           * Schema defining the parameters accepted by the function.
+           */
+          parameters: {
+            /**
+             * The type of the parameters object (usually 'object').
+             */
+            type: string;
+            /**
+             * List of required parameter names.
+             */
+            required?: string[];
+            /**
+             * Definitions of each parameter.
+             */
+            properties: {
+              [k: string]: {
+                /**
+                 * The data type of the parameter.
+                 */
+                type: string;
+                /**
+                 * A description of the expected parameter.
+                 */
+                description: string;
+              };
+            };
+          };
+        };
+      }
+  )[];
+  response_format?: Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode_3;
+  /**
+   * If true, a chat template is not applied and you must adhere to the specific model's expected formatting.
+   */
+  raw?: boolean;
+  /**
+   * If true, the response will be streamed back incrementally using SSE, Server Sent Events.
+   */
+  stream?: boolean;
+  /**
+   * The maximum number of tokens to generate in the response.
+   */
+  max_tokens?: number;
+  /**
+   * Controls the randomness of the output; higher values produce more random results.
+   */
+  temperature?: number;
+  /**
+   * Adjusts the creativity of the AI's responses by controlling how many possible words it considers. Lower values make outputs more predictable; higher values allow for more varied and creative responses.
+   */
+  top_p?: number;
+  /**
+   * Limits the AI to choose from the top 'k' most probable words. Lower values make responses more focused; higher values introduce more variety and potential surprises.
+   */
+  top_k?: number;
+  /**
+   * Random seed for reproducibility of the generation.
+   */
+  seed?: number;
+  /**
+   * Penalty for repeated tokens; higher values discourage repetition.
+   */
+  repetition_penalty?: number;
+  /**
+   * Decreases the likelihood of the model repeating the same lines verbatim.
+   */
+  frequency_penalty?: number;
+  /**
+   * Increases the likelihood of the model introducing new topics.
+   */
+  presence_penalty?: number;
 }
-declare abstract class Base_Ai_Cf_Deepgram_Nova_3 {
-  inputs: Ai_Cf_Deepgram_Nova_3_Input;
-  postProcessedOutputs: Ai_Cf_Deepgram_Nova_3_Output;
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_JSON_Mode_3 {
+  type?: "json_object" | "json_schema";
+  json_schema?: unknown;
 }
-type Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Input =
-  | {
+type Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Output =
+  | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Chat_Completion_Response
+  | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Text_Completion_Response
+  | string
+  | Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_AsyncResponse;
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Chat_Completion_Response {
+  /**
+   * Unique identifier for the completion
+   */
+  id?: string;
+  /**
+   * Object type identifier
+   */
+  object?: "chat.completion";
+  /**
+   * Unix timestamp of when the completion was created
+   */
+  created?: number;
+  /**
+   * Model used for the completion
+   */
+  model?: string;
+  /**
+   * List of completion choices
+   */
+  choices?: {
+    /**
+     * Index of the choice in the list
+     */
+    index?: number;
+    /**
+     * The message generated by the model
+     */
+    message?: {
       /**
-       * readable stream with audio data and content-type specified for that data
+       * Role of the message author
        */
-      audio: {
-        body: object;
-        contentType: string;
-      };
+      role: string;
       /**
-       * type of data PCM data that's sent to the inference server as raw array
+       * The content of the message
        */
-      dtype?: "uint8" | "float32" | "float64";
-    }
-  | {
+      content: string;
       /**
-       * base64 encoded audio data
+       * Internal reasoning content (if available)
        */
-      audio: string;
+      reasoning_content?: string;
       /**
-       * type of data PCM data that's sent to the inference server as raw array
+       * Tool calls made by the assistant
        */
-      dtype?: "uint8" | "float32" | "float64";
+      tool_calls?: {
+        /**
+         * Unique identifier for the tool call
+         */
+        id: string;
+        /**
+         * Type of tool call
+         */
+        type: "function";
+        function: {
+          /**
+           * Name of the function to call
+           */
+          name: string;
+          /**
+           * JSON string of arguments for the function
+           */
+          arguments: string;
+        };
+      }[];
     };
-interface Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Output {
+    /**
+     * Reason why the model stopped generating
+     */
+    finish_reason?: string;
+    /**
+     * Stop reason (may be null)
+     */
+    stop_reason?: string | null;
+    /**
+     * Log probabilities (if requested)
+     */
+    logprobs?: {} | null;
+  }[];
   /**
-   * if true, end-of-turn was detected
+   * Usage statistics for the inference request
    */
-  is_complete?: boolean;
+  usage?: {
+    /**
+     * Total number of tokens in input
+     */
+    prompt_tokens?: number;
+    /**
+     * Total number of tokens in output
+     */
+    completion_tokens?: number;
+    /**
+     * Total number of input and output tokens
+     */
+    total_tokens?: number;
+  };
   /**
-   * probability of the end-of-turn detection
+   * Log probabilities for the prompt (if requested)
    */
-  probability?: number;
-}
-declare abstract class Base_Ai_Cf_Pipecat_Ai_Smart_Turn_V2 {
-  inputs: Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Input;
-  postProcessedOutputs: Ai_Cf_Pipecat_Ai_Smart_Turn_V2_Output;
+  prompt_logprobs?: {} | null;
 }
-type Ai_Cf_Openai_Gpt_Oss_120B_Input =
-  | GPT_OSS_120B_Responses
-  | GPT_OSS_120B_Responses_Async;
-interface GPT_OSS_120B_Responses {
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Text_Completion_Response {
+  /**
+   * Unique identifier for the completion
+   */
+  id?: string;
+  /**
+   * Object type identifier
+   */
+  object?: "text_completion";
+  /**
+   * Unix timestamp of when the completion was created
+   */
+  created?: number;
+  /**
+   * Model used for the completion
+   */
+  model?: string;
   /**
-   * Responses API Input messages. Refer to OpenAI Responses API docs to learn more about supported content types
+   * List of completion choices
    */
-  input: string | unknown[];
-  reasoning?: {
+  choices?: {
     /**
-     * Constrains effort on reasoning for reasoning models. Currently supported values are low, medium, and high. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response.
+     * Index of the choice in the list
      */
-    effort?: "low" | "medium" | "high";
+    index: number;
     /**
-     * A summary of the reasoning performed by the model. This can be useful for debugging and understanding the model's reasoning process. One of auto, concise, or detailed.
+     * The generated text completion
      */
-    summary?: "auto" | "concise" | "detailed";
-  };
-}
-interface GPT_OSS_120B_Responses_Async {
-  requests: {
+    text: string;
     /**
-     * Responses API Input messages. Refer to OpenAI Responses API docs to learn more about supported content types
+     * Reason why the model stopped generating
      */
-    input: string | unknown[];
-    reasoning?: {
-      /**
-       * Constrains effort on reasoning for reasoning models. Currently supported values are low, medium, and high. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response.
-       */
-      effort?: "low" | "medium" | "high";
-      /**
-       * A summary of the reasoning performed by the model. This can be useful for debugging and understanding the model's reasoning process. One of auto, concise, or detailed.
-       */
-      summary?: "auto" | "concise" | "detailed";
-    };
+    finish_reason: string;
+    /**
+     * Stop reason (may be null)
+     */
+    stop_reason?: string | null;
+    /**
+     * Log probabilities (if requested)
+     */
+    logprobs?: {} | null;
+    /**
+     * Log probabilities for the prompt (if requested)
+     */
+    prompt_logprobs?: {} | null;
   }[];
-}
-type Ai_Cf_Openai_Gpt_Oss_120B_Output = {} | (string & NonNullable<unknown>);
-declare abstract class Base_Ai_Cf_Openai_Gpt_Oss_120B {
-  inputs: Ai_Cf_Openai_Gpt_Oss_120B_Input;
-  postProcessedOutputs: Ai_Cf_Openai_Gpt_Oss_120B_Output;
-}
-type Ai_Cf_Openai_Gpt_Oss_20B_Input =
-  | GPT_OSS_20B_Responses
-  | GPT_OSS_20B_Responses_Async;
-interface GPT_OSS_20B_Responses {
   /**
-   * Responses API Input messages. Refer to OpenAI Responses API docs to learn more about supported content types
+   * Usage statistics for the inference request
    */
-  input: string | unknown[];
-  reasoning?: {
+  usage?: {
     /**
-     * Constrains effort on reasoning for reasoning models. Currently supported values are low, medium, and high. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response.
+     * Total number of tokens in input
      */
-    effort?: "low" | "medium" | "high";
+    prompt_tokens?: number;
     /**
-     * A summary of the reasoning performed by the model. This can be useful for debugging and understanding the model's reasoning process. One of auto, concise, or detailed.
+     * Total number of tokens in output
      */
-    summary?: "auto" | "concise" | "detailed";
-  };
-}
-interface GPT_OSS_20B_Responses_Async {
-  requests: {
+    completion_tokens?: number;
     /**
-     * Responses API Input messages. Refer to OpenAI Responses API docs to learn more about supported content types
+     * Total number of input and output tokens
      */
-    input: string | unknown[];
-    reasoning?: {
-      /**
-       * Constrains effort on reasoning for reasoning models. Currently supported values are low, medium, and high. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response.
-       */
-      effort?: "low" | "medium" | "high";
-      /**
-       * A summary of the reasoning performed by the model. This can be useful for debugging and understanding the model's reasoning process. One of auto, concise, or detailed.
-       */
-      summary?: "auto" | "concise" | "detailed";
-    };
-  }[];
+    total_tokens?: number;
+  };
 }
-type Ai_Cf_Openai_Gpt_Oss_20B_Output = {} | (string & NonNullable<unknown>);
-declare abstract class Base_Ai_Cf_Openai_Gpt_Oss_20B {
-  inputs: Ai_Cf_Openai_Gpt_Oss_20B_Input;
-  postProcessedOutputs: Ai_Cf_Openai_Gpt_Oss_20B_Output;
+interface Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_AsyncResponse {
+  /**
+   * The async request id that can be used to obtain the results.
+   */
+  request_id?: string;
 }
-interface Ai_Cf_Leonardo_Phoenix_1_0_Input {
+declare abstract class Base_Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It {
+  inputs: Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Input;
+  postProcessedOutputs: Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It_Output;
+}
+interface Ai_Cf_Pfnet_Plamo_Embedding_1B_Input {
   /**
-   * A text description of the image you want to generate.
+   * Input text to embed. Can be a single string or a list of strings.
    */
-  prompt: string;
+  text: string | string[];
+}
+interface Ai_Cf_Pfnet_Plamo_Embedding_1B_Output {
   /**
-   * Controls how closely the generated image should adhere to the prompt; higher values make the image more aligned with the prompt
+   * Embedding vectors, where each vector is a list of floats.
    */
-  guidance?: number;
+  data: number[][];
   /**
-   * Random seed for reproducibility of the image generation
+   * Shape of the embedding data as [number_of_embeddings, embedding_dimension].
+   *
+   * @minItems 2
+   * @maxItems 2
    */
-  seed?: number;
+  shape: [number, number];
+}
+declare abstract class Base_Ai_Cf_Pfnet_Plamo_Embedding_1B {
+  inputs: Ai_Cf_Pfnet_Plamo_Embedding_1B_Input;
+  postProcessedOutputs: Ai_Cf_Pfnet_Plamo_Embedding_1B_Output;
+}
+interface Ai_Cf_Deepgram_Flux_Input {
   /**
-   * The height of the generated image in pixels
+   * Encoding of the audio stream. Currently only supports raw signed little-endian 16-bit PCM.
    */
-  height?: number;
+  encoding: "linear16";
   /**
-   * The width of the generated image in pixels
+   * Sample rate of the audio stream in Hz.
    */
-  width?: number;
+  sample_rate: string;
   /**
-   * The number of diffusion steps; higher values can improve quality but take longer
+   * End-of-turn confidence required to fire an eager end-of-turn event. When set, enables EagerEndOfTurn and TurnResumed events. Valid Values 0.3 - 0.9.
    */
-  num_steps?: number;
+  eager_eot_threshold?: string;
   /**
-   * Specify what to exclude from the generated images
+   * End-of-turn confidence required to finish a turn. Valid Values 0.5 - 0.9.
    */
-  negative_prompt?: string;
+  eot_threshold?: string;
+  /**
+   * A turn will be finished when this much time has passed after speech, regardless of EOT confidence.
+   */
+  eot_timeout_ms?: string;
+  /**
+   * Keyterm prompting can improve recognition of specialized terminology. Pass multiple keyterm query parameters to boost multiple keyterms.
+   */
+  keyterm?: string;
+  /**
+   * Opts out requests from the Deepgram Model Improvement Program. Refer to Deepgram Docs for pricing impacts before setting this to true. https://dpgr.am/deepgram-mip
+   */
+  mip_opt_out?: "true" | "false";
+  /**
+   * Label your requests for the purpose of identification during usage reporting
+   */
+  tag?: string;
 }
 /**
- * The generated image in JPEG format
+ * Output will be returned as websocket messages.
  */
-type Ai_Cf_Leonardo_Phoenix_1_0_Output = string;
-declare abstract class Base_Ai_Cf_Leonardo_Phoenix_1_0 {
-  inputs: Ai_Cf_Leonardo_Phoenix_1_0_Input;
-  postProcessedOutputs: Ai_Cf_Leonardo_Phoenix_1_0_Output;
-}
-interface Ai_Cf_Leonardo_Lucid_Origin_Input {
+interface Ai_Cf_Deepgram_Flux_Output {
   /**
-   * A text description of the image you want to generate.
+   * The unique identifier of the request (uuid)
    */
-  prompt: string;
+  request_id?: string;
   /**
-   * Controls how closely the generated image should adhere to the prompt; higher values make the image more aligned with the prompt
+   * Starts at 0 and increments for each message the server sends to the client.
    */
-  guidance?: number;
+  sequence_id?: number;
   /**
-   * Random seed for reproducibility of the image generation
+   * The type of event being reported.
    */
-  seed?: number;
+  event?:
+    | "Update"
+    | "StartOfTurn"
+    | "EagerEndOfTurn"
+    | "TurnResumed"
+    | "EndOfTurn";
   /**
-   * The height of the generated image in pixels
+   * The index of the current turn
    */
-  height?: number;
+  turn_index?: number;
   /**
-   * The width of the generated image in pixels
+   * Start time in seconds of the audio range that was transcribed
    */
-  width?: number;
+  audio_window_start?: number;
   /**
-   * The number of diffusion steps; higher values can improve quality but take longer
+   * End time in seconds of the audio range that was transcribed
    */
-  num_steps?: number;
+  audio_window_end?: number;
   /**
-   * The number of diffusion steps; higher values can improve quality but take longer
+   * Text that was said over the course of the current turn
    */
-  steps?: number;
-}
-interface Ai_Cf_Leonardo_Lucid_Origin_Output {
+  transcript?: string;
   /**
-   * The generated image in Base64 format.
+   * The words in the transcript
    */
-  image?: string;
+  words?: {
+    /**
+     * The individual punctuated, properly-cased word from the transcript
+     */
+    word: string;
+    /**
+     * Confidence that this word was transcribed correctly
+     */
+    confidence: number;
+  }[];
+  /**
+   * Confidence that no more speech is coming in this turn
+   */
+  end_of_turn_confidence?: number;
 }
-declare abstract class Base_Ai_Cf_Leonardo_Lucid_Origin {
-  inputs: Ai_Cf_Leonardo_Lucid_Origin_Input;
-  postProcessedOutputs: Ai_Cf_Leonardo_Lucid_Origin_Output;
+declare abstract class Base_Ai_Cf_Deepgram_Flux {
+  inputs: Ai_Cf_Deepgram_Flux_Input;
+  postProcessedOutputs: Ai_Cf_Deepgram_Flux_Output;
 }
-interface Ai_Cf_Deepgram_Aura_1_Input {
+interface Ai_Cf_Deepgram_Aura_2_En_Input {
   /**
    * Speaker used to produce the audio.
    */
   speaker?:
-    | "angus"
-    | "asteria"
+    | "amalthea"
+    | "andromeda"
+    | "apollo"
     | "arcas"
-    | "orion"
-    | "orpheus"
+    | "aries"
+    | "asteria"
     | "athena"
-    | "luna"
-    | "zeus"
-    | "perseus"
-    | "helios"
+    | "atlas"
+    | "aurora"
+    | "callista"
+    | "cora"
+    | "cordelia"
+    | "delia"
+    | "draco"
+    | "electra"
+    | "harmonia"
+    | "helena"
     | "hera"
-    | "stella";
+    | "hermes"
+    | "hyperion"
+    | "iris"
+    | "janus"
+    | "juno"
+    | "jupiter"
+    | "luna"
+    | "mars"
+    | "minerva"
+    | "neptune"
+    | "odysseus"
+    | "ophelia"
+    | "orion"
+    | "orpheus"
+    | "pandora"
+    | "phoebe"
+    | "pluto"
+    | "saturn"
+    | "thalia"
+    | "theia"
+    | "vesta"
+    | "zeus";
   /**
    * Encoding of the output audio.
    */
@@ -7277,10 +9141,54 @@ interface Ai_Cf_Deepgram_Aura_1_Input {
 /**
  * The generated audio in MP3 format
  */
-type Ai_Cf_Deepgram_Aura_1_Output = string;
-declare abstract class Base_Ai_Cf_Deepgram_Aura_1 {
-  inputs: Ai_Cf_Deepgram_Aura_1_Input;
-  postProcessedOutputs: Ai_Cf_Deepgram_Aura_1_Output;
+type Ai_Cf_Deepgram_Aura_2_En_Output = string;
+declare abstract class Base_Ai_Cf_Deepgram_Aura_2_En {
+  inputs: Ai_Cf_Deepgram_Aura_2_En_Input;
+  postProcessedOutputs: Ai_Cf_Deepgram_Aura_2_En_Output;
+}
+interface Ai_Cf_Deepgram_Aura_2_Es_Input {
+  /**
+   * Speaker used to produce the audio.
+   */
+  speaker?:
+    | "sirio"
+    | "nestor"
+    | "carina"
+    | "celeste"
+    | "alvaro"
+    | "diana"
+    | "aquila"
+    | "selena"
+    | "estrella"
+    | "javier";
+  /**
+   * Encoding of the output audio.
+   */
+  encoding?: "linear16" | "flac" | "mulaw" | "alaw" | "mp3" | "opus" | "aac";
+  /**
+   * Container specifies the file format wrapper for the output audio. The available options depend on the encoding type..
+   */
+  container?: "none" | "wav" | "ogg";
+  /**
+   * The text content to be converted to speech
+   */
+  text: string;
+  /**
+   * Sample Rate specifies the sample rate for the output audio. Based on the encoding, different sample rates are supported. For some encodings, the sample rate is not configurable
+   */
+  sample_rate?: number;
+  /**
+   * The bitrate of the audio in bits per second. Choose from predefined ranges or specific values based on the encoding type.
+   */
+  bit_rate?: number;
+}
+/**
+ * The generated audio in MP3 format
+ */
+type Ai_Cf_Deepgram_Aura_2_Es_Output = string;
+declare abstract class Base_Ai_Cf_Deepgram_Aura_2_Es {
+  inputs: Ai_Cf_Deepgram_Aura_2_Es_Input;
+  postProcessedOutputs: Ai_Cf_Deepgram_Aura_2_Es_Output;
 }
 interface AiModels {
   "@cf/huggingface/distilbert-sst-2-int8": BaseAiTextClassification;
@@ -7325,12 +9233,12 @@ interface AiModels {
   "@cf/meta/llama-3-8b-instruct": BaseAiTextGeneration;
   "@cf/fblgit/una-cybertron-7b-v2-bf16": BaseAiTextGeneration;
   "@cf/meta/llama-3-8b-instruct-awq": BaseAiTextGeneration;
-  "@hf/meta-llama/meta-llama-3-8b-instruct": BaseAiTextGeneration;
   "@cf/meta/llama-3.1-8b-instruct-fp8": BaseAiTextGeneration;
   "@cf/meta/llama-3.1-8b-instruct-awq": BaseAiTextGeneration;
   "@cf/meta/llama-3.2-3b-instruct": BaseAiTextGeneration;
   "@cf/meta/llama-3.2-1b-instruct": BaseAiTextGeneration;
   "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b": BaseAiTextGeneration;
+  "@cf/ibm-granite/granite-4.0-h-micro": BaseAiTextGeneration;
   "@cf/facebook/bart-large-cnn": BaseAiSummarization;
   "@cf/llava-hf/llava-1.5-7b-hf": BaseAiImageToText;
   "@cf/baai/bge-base-en-v1.5": Base_Ai_Cf_Baai_Bge_Base_En_V1_5;
@@ -7352,13 +9260,21 @@ interface AiModels {
   "@cf/mistralai/mistral-small-3.1-24b-instruct": Base_Ai_Cf_Mistralai_Mistral_Small_3_1_24B_Instruct;
   "@cf/google/gemma-3-12b-it": Base_Ai_Cf_Google_Gemma_3_12B_It;
   "@cf/meta/llama-4-scout-17b-16e-instruct": Base_Ai_Cf_Meta_Llama_4_Scout_17B_16E_Instruct;
+  "@cf/qwen/qwen3-30b-a3b-fp8": Base_Ai_Cf_Qwen_Qwen3_30B_A3B_Fp8;
   "@cf/deepgram/nova-3": Base_Ai_Cf_Deepgram_Nova_3;
+  "@cf/qwen/qwen3-embedding-0.6b": Base_Ai_Cf_Qwen_Qwen3_Embedding_0_6B;
   "@cf/pipecat-ai/smart-turn-v2": Base_Ai_Cf_Pipecat_Ai_Smart_Turn_V2;
   "@cf/openai/gpt-oss-120b": Base_Ai_Cf_Openai_Gpt_Oss_120B;
   "@cf/openai/gpt-oss-20b": Base_Ai_Cf_Openai_Gpt_Oss_20B;
   "@cf/leonardo/phoenix-1.0": Base_Ai_Cf_Leonardo_Phoenix_1_0;
   "@cf/leonardo/lucid-origin": Base_Ai_Cf_Leonardo_Lucid_Origin;
   "@cf/deepgram/aura-1": Base_Ai_Cf_Deepgram_Aura_1;
+  "@cf/ai4bharat/indictrans2-en-indic-1B": Base_Ai_Cf_Ai4Bharat_Indictrans2_En_Indic_1B;
+  "@cf/aisingapore/gemma-sea-lion-v4-27b-it": Base_Ai_Cf_Aisingapore_Gemma_Sea_Lion_V4_27B_It;
+  "@cf/pfnet/plamo-embedding-1b": Base_Ai_Cf_Pfnet_Plamo_Embedding_1B;
+  "@cf/deepgram/flux": Base_Ai_Cf_Deepgram_Flux;
+  "@cf/deepgram/aura-2-en": Base_Ai_Cf_Deepgram_Aura_2_En;
+  "@cf/deepgram/aura-2-es": Base_Ai_Cf_Deepgram_Aura_2_Es;
 }
 type AiOptions = {
   /**
@@ -7370,6 +9286,16 @@ type AiOptions = {
    * Establish websocket connections, only works for supported models
    */
   websocket?: boolean;
+  /**
+   * Tag your requests to group and view them in Cloudflare dashboard.
+   *
+   * Rules:
+   * Tags must only contain letters, numbers, and the symbols: : - . / @
+   * Each tag can have maximum 50 characters.
+   * Maximum 5 tags are allowed each request.
+   * Duplicate tags will removed.
+   */
+  tags: string[];
   gateway?: GatewayOptions;
   returnRawResponse?: boolean;
   prefix?: string;
@@ -9457,12 +11383,13 @@ declare namespace Rpc {
   export type Provider<
     T extends object,
     Reserved extends string = never,
-  > = MaybeCallableProvider<T> & {
-    [K in Exclude<
-      keyof T,
-      Reserved | symbol | keyof StubBase<never>
-    >]: MethodOrProperty<T[K]>;
-  };
+  > = MaybeCallableProvider<T> &
+    Pick<
+      {
+        [K in keyof T]: MethodOrProperty<T[K]>;
+      },
+      Exclude<keyof T, Reserved | symbol | keyof StubBase<never>>
+    >;
 }
 declare namespace Cloudflare {
   // Type of `env`.