@openrouter/sdk 0.12.35 → 0.13.7

package/esm/models/chatrequest.d.ts

@@ -15,12 +15,15 @@ import { ChatToolChoice, ChatToolChoice$Outbound } from "./chattoolchoice.js";
 import { ContextCompressionPlugin, ContextCompressionPlugin$Outbound } from "./contextcompressionplugin.js";
 import { FileParserPlugin, FileParserPlugin$Outbound } from "./fileparserplugin.js";
 import { FormatJsonObjectConfig, FormatJsonObjectConfig$Outbound } from "./formatjsonobjectconfig.js";
+import { FusionPlugin, FusionPlugin$Outbound } from "./fusionplugin.js";
 import { ImageConfig, ImageConfig$Outbound } from "./imageconfig.js";
 import { ModerationPlugin, ModerationPlugin$Outbound } from "./moderationplugin.js";
 import { ParetoRouterPlugin, ParetoRouterPlugin$Outbound } from "./paretorouterplugin.js";
 import { ProviderPreferences, ProviderPreferences$Outbound } from "./providerpreferences.js";
 import { ResponseHealingPlugin, ResponseHealingPlugin$Outbound } from "./responsehealingplugin.js";
+import { StopServerToolsWhenCondition, StopServerToolsWhenCondition$Outbound } from "./stopservertoolswhencondition.js";
 import { TraceConfig, TraceConfig$Outbound } from "./traceconfig.js";
+import { WebFetchPlugin, WebFetchPlugin$Outbound } from "./webfetchplugin.js";
 import { WebSearchPlugin, WebSearchPlugin$Outbound } from "./websearchplugin.js";
 export declare const Modality: {
     readonly Text: "text";
@@ -28,11 +31,11 @@ export declare const Modality: {
     readonly Audio: "audio";
 };
 export type Modality = OpenEnum<typeof Modality>;
-export type ChatRequestPlugin = AutoRouterPlugin | ContextCompressionPlugin | FileParserPlugin | ModerationPlugin | ParetoRouterPlugin | ResponseHealingPlugin | WebSearchPlugin;
+export type ChatRequestPlugin = AutoRouterPlugin | ContextCompressionPlugin | FileParserPlugin | FusionPlugin | ModerationPlugin | ParetoRouterPlugin | ResponseHealingPlugin | WebSearchPlugin | WebFetchPlugin;
 /**
  * Constrains effort on reasoning for reasoning models
  */
-export declare const Effort: {
+export declare const ChatRequestEffort: {
     readonly Xhigh: "xhigh";
     readonly High: "high";
     readonly Medium: "medium";
@@ -43,17 +46,32 @@ export declare const Effort: {
 /**
  * Constrains effort on reasoning for reasoning models
  */
-export type Effort = OpenEnum<typeof Effort>;
+export type ChatRequestEffort = OpenEnum<typeof ChatRequestEffort>;
 /**
  * Configuration options for reasoning models
  */
-export type Reasoning = {
+export type ChatRequestReasoning = {
     /**
      * Constrains effort on reasoning for reasoning models
      */
-    effort?: Effort | null | undefined;
+    effort?: ChatRequestEffort | null | undefined;
     summary?: ChatReasoningSummaryVerbosityEnum | null | undefined;
 };
+/**
+ * Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
+ */
+export declare const ChatRequestReasoningEffort: {
+    readonly Xhigh: "xhigh";
+    readonly High: "high";
+    readonly Medium: "medium";
+    readonly Low: "low";
+    readonly Minimal: "minimal";
+    readonly None: "none";
+};
+/**
+ * Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
+ */
+export type ChatRequestReasoningEffort = OpenEnum<typeof ChatRequestReasoningEffort>;
 /**
  * Response format configuration
  */
@@ -80,6 +98,9 @@ export type Stop = string | Array<string> | any;
  * Chat completion request parameters
  */
 export type ChatRequest = {
+    /**
+     * Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.
+     */
     cacheControl?: AnthropicCacheControlDirective | undefined;
     /**
      * Debug options for inspecting request transformations (streaming only)
@@ -123,6 +144,10 @@ export type ChatRequest = {
     metadata?: {
         [k: string]: string;
     } | undefined;
+    /**
+     * Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter.
+     */
+    minP?: number | null | undefined;
     /**
      * Output modalities for the response. Supported values are "text", "image", and "audio".
      */
@@ -142,7 +167,7 @@ export type ChatRequest = {
     /**
      * Plugins you want to enable for this request, including their settings.
      */
-    plugins?: Array<AutoRouterPlugin | ContextCompressionPlugin | FileParserPlugin | ModerationPlugin | ParetoRouterPlugin | ResponseHealingPlugin | WebSearchPlugin> | undefined;
+    plugins?: Array<AutoRouterPlugin | ContextCompressionPlugin | FileParserPlugin | FusionPlugin | ModerationPlugin | ParetoRouterPlugin | ResponseHealingPlugin | WebSearchPlugin | WebFetchPlugin> | undefined;
     /**
      * Presence penalty (-2.0 to 2.0)
      */
@@ -154,7 +179,15 @@ export type ChatRequest = {
     /**
      * Configuration options for reasoning models
      */
-    reasoning?: Reasoning | undefined;
+    reasoning?: ChatRequestReasoning | undefined;
+    /**
+     * Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
+     */
+    reasoningEffort?: ChatRequestReasoningEffort | null | undefined;
+    /**
+     * Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter.
+     */
+    repetitionPenalty?: number | null | undefined;
     /**
      * Response format configuration
      */
@@ -168,13 +201,17 @@ export type ChatRequest = {
      */
     serviceTier?: ChatRequestServiceTier | null | undefined;
     /**
-     * A unique identifier for grouping related requests (e.g., a conversation or agent workflow) for observability. If provided in both the request body and the x-session-id header, the body value takes precedence. Maximum of 256 characters.
+     * A unique identifier for grouping related requests (e.g., a conversation or agent workflow). When provided, OpenRouter uses it as the sticky routing key, routing all requests in the session to the same provider to maximize prompt cache hits. Also used for observability grouping. If provided in both the request body and the x-session-id header, the body value takes precedence. Maximum of 256 characters.
      */
     sessionId?: string | undefined;
     /**
      * Stop sequences (up to 4)
      */
     stop?: string | Array<string> | any | null | undefined;
+    /**
+     * Stop conditions for the server-tool agent loop. Any condition firing halts the loop (OR logic). When set, this overrides `max_tool_calls`.
+     */
+    stopServerToolsWhen?: Array<StopServerToolsWhenCondition> | undefined;
     /**
      * Enable streaming response
      */
@@ -195,6 +232,14 @@ export type ChatRequest = {
      * Available tools for function calling
      */
     tools?: Array<ChatFunctionTool> | undefined;
+    /**
+     * Consider only tokens with "sufficiently high" probabilities based on the probability of the most likely token. Not all providers support this parameter.
+     */
+    topA?: number | null | undefined;
+    /**
+     * Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter.
+     */
+    topK?: number | null | undefined;
     /**
      * Number of top log probabilities to return (0-20)
      */
@@ -215,20 +260,22 @@ export type ChatRequest = {
 /** @internal */
 export declare const Modality$outboundSchema: z.ZodType<string, Modality>;
 /** @internal */
-export type ChatRequestPlugin$Outbound = AutoRouterPlugin$Outbound | ContextCompressionPlugin$Outbound | FileParserPlugin$Outbound | ModerationPlugin$Outbound | ParetoRouterPlugin$Outbound | ResponseHealingPlugin$Outbound | WebSearchPlugin$Outbound;
+export type ChatRequestPlugin$Outbound = AutoRouterPlugin$Outbound | ContextCompressionPlugin$Outbound | FileParserPlugin$Outbound | FusionPlugin$Outbound | ModerationPlugin$Outbound | ParetoRouterPlugin$Outbound | ResponseHealingPlugin$Outbound | WebSearchPlugin$Outbound | WebFetchPlugin$Outbound;
 /** @internal */
 export declare const ChatRequestPlugin$outboundSchema: z.ZodType<ChatRequestPlugin$Outbound, ChatRequestPlugin>;
 export declare function chatRequestPluginToJSON(chatRequestPlugin: ChatRequestPlugin): string;
 /** @internal */
-export declare const Effort$outboundSchema: z.ZodType<string, Effort>;
+export declare const ChatRequestEffort$outboundSchema: z.ZodType<string, ChatRequestEffort>;
 /** @internal */
-export type Reasoning$Outbound = {
+export type ChatRequestReasoning$Outbound = {
     effort?: string | null | undefined;
     summary?: string | null | undefined;
 };
 /** @internal */
-export declare const Reasoning$outboundSchema: z.ZodType<Reasoning$Outbound, Reasoning>;
-export declare function reasoningToJSON(reasoning: Reasoning): string;
+export declare const ChatRequestReasoning$outboundSchema: z.ZodType<ChatRequestReasoning$Outbound, ChatRequestReasoning>;
+export declare function chatRequestReasoningToJSON(chatRequestReasoning: ChatRequestReasoning): string;
+/** @internal */
+export declare const ChatRequestReasoningEffort$outboundSchema: z.ZodType<string, ChatRequestReasoningEffort>;
 /** @internal */
 export type ResponseFormat$Outbound = ChatFormatGrammarConfig$Outbound | FormatJsonObjectConfig$Outbound | ChatFormatJsonSchemaConfig$Outbound | ChatFormatPythonConfig$Outbound | ChatFormatTextConfig$Outbound;
 /** @internal */
@@ -259,24 +306,30 @@ export type ChatRequest$Outbound = {
     metadata?: {
         [k: string]: string;
     } | undefined;
+    min_p?: number | null | undefined;
     modalities?: Array<string> | undefined;
     model?: string | undefined;
     models?: Array<string> | undefined;
     parallel_tool_calls?: boolean | null | undefined;
-    plugins?: Array<AutoRouterPlugin$Outbound | ContextCompressionPlugin$Outbound | FileParserPlugin$Outbound | ModerationPlugin$Outbound | ParetoRouterPlugin$Outbound | ResponseHealingPlugin$Outbound | WebSearchPlugin$Outbound> | undefined;
+    plugins?: Array<AutoRouterPlugin$Outbound | ContextCompressionPlugin$Outbound | FileParserPlugin$Outbound | FusionPlugin$Outbound | ModerationPlugin$Outbound | ParetoRouterPlugin$Outbound | ResponseHealingPlugin$Outbound | WebSearchPlugin$Outbound | WebFetchPlugin$Outbound> | undefined;
     presence_penalty?: number | null | undefined;
     provider?: ProviderPreferences$Outbound | null | undefined;
-    reasoning?: Reasoning$Outbound | undefined;
+    reasoning?: ChatRequestReasoning$Outbound | undefined;
+    reasoning_effort?: string | null | undefined;
+    repetition_penalty?: number | null | undefined;
     response_format?: ChatFormatGrammarConfig$Outbound | FormatJsonObjectConfig$Outbound | ChatFormatJsonSchemaConfig$Outbound | ChatFormatPythonConfig$Outbound | ChatFormatTextConfig$Outbound | undefined;
     seed?: number | null | undefined;
     service_tier?: string | null | undefined;
     session_id?: string | undefined;
     stop?: string | Array<string> | any | null | undefined;
+    stop_server_tools_when?: Array<StopServerToolsWhenCondition$Outbound> | undefined;
     stream: boolean;
     stream_options?: ChatStreamOptions$Outbound | null | undefined;
     temperature?: number | null | undefined;
     tool_choice?: ChatToolChoice$Outbound | undefined;
     tools?: Array<ChatFunctionTool$Outbound> | undefined;
+    top_a?: number | null | undefined;
+    top_k?: number | null | undefined;
     top_logprobs?: number | null | undefined;
     top_p?: number | null | undefined;
     trace?: TraceConfig$Outbound | undefined;

package/esm/models/chatrequest.js

@@ -20,12 +20,15 @@ import { ChatToolChoice$outboundSchema, } from "./chattoolchoice.js";
 import { ContextCompressionPlugin$outboundSchema, } from "./contextcompressionplugin.js";
 import { FileParserPlugin$outboundSchema, } from "./fileparserplugin.js";
 import { FormatJsonObjectConfig$outboundSchema, } from "./formatjsonobjectconfig.js";
+import { FusionPlugin$outboundSchema, } from "./fusionplugin.js";
 import { ImageConfig$outboundSchema, } from "./imageconfig.js";
 import { ModerationPlugin$outboundSchema, } from "./moderationplugin.js";
 import { ParetoRouterPlugin$outboundSchema, } from "./paretorouterplugin.js";
 import { ProviderPreferences$outboundSchema, } from "./providerpreferences.js";
 import { ResponseHealingPlugin$outboundSchema, } from "./responsehealingplugin.js";
+import { StopServerToolsWhenCondition$outboundSchema, } from "./stopservertoolswhencondition.js";
 import { TraceConfig$outboundSchema, } from "./traceconfig.js";
+import { WebFetchPlugin$outboundSchema, } from "./webfetchplugin.js";
 import { WebSearchPlugin$outboundSchema, } from "./websearchplugin.js";
 export const Modality = {
     Text: "text",
@@ -35,7 +38,18 @@ export const Modality = {
 /**
  * Constrains effort on reasoning for reasoning models
  */
-export const Effort = {
+export const ChatRequestEffort = {
+    Xhigh: "xhigh",
+    High: "high",
+    Medium: "medium",
+    Low: "low",
+    Minimal: "minimal",
+    None: "none",
+};
+/**
+ * Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
+ */
+export const ChatRequestReasoningEffort = {
     Xhigh: "xhigh",
     High: "high",
     Medium: "medium",
@@ -61,27 +75,30 @@ export const ChatRequestPlugin$outboundSchema = z.union([
     AutoRouterPlugin$outboundSchema,
     ContextCompressionPlugin$outboundSchema,
     FileParserPlugin$outboundSchema,
+    FusionPlugin$outboundSchema,
     ModerationPlugin$outboundSchema,
     ParetoRouterPlugin$outboundSchema,
     ResponseHealingPlugin$outboundSchema,
     WebSearchPlugin$outboundSchema,
+    WebFetchPlugin$outboundSchema,
 ]);
 export function chatRequestPluginToJSON(chatRequestPlugin) {
     return JSON.stringify(ChatRequestPlugin$outboundSchema.parse(chatRequestPlugin));
 }
 /** @internal */
-export const Effort$outboundSchema = openEnums
-    .outboundSchema(Effort);
+export const ChatRequestEffort$outboundSchema = openEnums.outboundSchema(ChatRequestEffort);
 /** @internal */
-export const Reasoning$outboundSchema = z.object({
-    effort: z.nullable(Effort$outboundSchema).optional(),
+export const ChatRequestReasoning$outboundSchema = z.object({
+    effort: z.nullable(ChatRequestEffort$outboundSchema).optional(),
     summary: z.nullable(ChatReasoningSummaryVerbosityEnum$outboundSchema)
         .optional(),
 });
-export function reasoningToJSON(reasoning) {
-    return JSON.stringify(Reasoning$outboundSchema.parse(reasoning));
+export function chatRequestReasoningToJSON(chatRequestReasoning) {
+    return JSON.stringify(ChatRequestReasoning$outboundSchema.parse(chatRequestReasoning));
 }
 /** @internal */
+export const ChatRequestReasoningEffort$outboundSchema = openEnums.outboundSchema(ChatRequestReasoningEffort);
+/** @internal */
 export const ResponseFormat$outboundSchema = z.union([
     ChatFormatGrammarConfig$outboundSchema,
     FormatJsonObjectConfig$outboundSchema,
@@ -115,6 +132,7 @@ export const ChatRequest$outboundSchema = z.object({
     maxTokens: z.nullable(z.int()).optional(),
     messages: z.array(ChatMessages$outboundSchema),
     metadata: z.record(z.string(), z.string()).optional(),
+    minP: z.nullable(z.number()).optional(),
     modalities: z.array(Modality$outboundSchema).optional(),
     model: z.string().optional(),
     models: z.array(z.string()).optional(),
@@ -123,14 +141,19 @@ export const ChatRequest$outboundSchema = z.object({
         AutoRouterPlugin$outboundSchema,
         ContextCompressionPlugin$outboundSchema,
         FileParserPlugin$outboundSchema,
+        FusionPlugin$outboundSchema,
         ModerationPlugin$outboundSchema,
         ParetoRouterPlugin$outboundSchema,
         ResponseHealingPlugin$outboundSchema,
         WebSearchPlugin$outboundSchema,
+        WebFetchPlugin$outboundSchema,
     ])).optional(),
     presencePenalty: z.nullable(z.number()).optional(),
     provider: z.nullable(ProviderPreferences$outboundSchema).optional(),
-    reasoning: z.lazy(() => Reasoning$outboundSchema).optional(),
+    reasoning: z.lazy(() => ChatRequestReasoning$outboundSchema).optional(),
+    reasoningEffort: z.nullable(ChatRequestReasoningEffort$outboundSchema)
+        .optional(),
+    repetitionPenalty: z.nullable(z.number()).optional(),
     responseFormat: z.union([
         ChatFormatGrammarConfig$outboundSchema,
         FormatJsonObjectConfig$outboundSchema,
@@ -143,11 +166,15 @@ export const ChatRequest$outboundSchema = z.object({
     sessionId: z.string().optional(),
     stop: z.nullable(z.union([z.string(), z.array(z.string()), z.any()]))
         .optional(),
+    stopServerToolsWhen: z.array(StopServerToolsWhenCondition$outboundSchema)
+        .optional(),
     stream: z.boolean().default(false),
     streamOptions: z.nullable(ChatStreamOptions$outboundSchema).optional(),
     temperature: z.nullable(z.number()).optional(),
     toolChoice: ChatToolChoice$outboundSchema.optional(),
     tools: z.array(ChatFunctionTool$outboundSchema).optional(),
+    topA: z.nullable(z.number()).optional(),
+    topK: z.nullable(z.int()).optional(),
     topLogprobs: z.nullable(z.int()).optional(),
     topP: z.nullable(z.number()).optional(),
     trace: TraceConfig$outboundSchema.optional(),
@@ -160,13 +187,19 @@ export const ChatRequest$outboundSchema = z.object({
         logitBias: "logit_bias",
         maxCompletionTokens: "max_completion_tokens",
         maxTokens: "max_tokens",
+        minP: "min_p",
         parallelToolCalls: "parallel_tool_calls",
         presencePenalty: "presence_penalty",
+        reasoningEffort: "reasoning_effort",
+        repetitionPenalty: "repetition_penalty",
         responseFormat: "response_format",
         serviceTier: "service_tier",
         sessionId: "session_id",
+        stopServerToolsWhen: "stop_server_tools_when",
         streamOptions: "stream_options",
         toolChoice: "tool_choice",
+        topA: "top_a",
+        topK: "top_k",
         topLogprobs: "top_logprobs",
         topP: "top_p",
     });

package/esm/models/chatservertoolchoice.d.ts

@@ -0,0 +1,18 @@
+import * as z from "zod/v4";
+/**
+ * OpenRouter extension: force a specific server tool by naming it directly in `tool_choice.type` instead of wrapping it in `{ type: "function", function: { name } }`.
+ */
+export type ChatServerToolChoice = {
+    /**
+     * OpenRouter server-tool type to force (e.g. `openrouter:web_search`, `web_search`, `web_search_preview`).
+     */
+    type: string;
+};
+/** @internal */
+export type ChatServerToolChoice$Outbound = {
+    type: string;
+};
+/** @internal */
+export declare const ChatServerToolChoice$outboundSchema: z.ZodType<ChatServerToolChoice$Outbound, ChatServerToolChoice>;
+export declare function chatServerToolChoiceToJSON(chatServerToolChoice: ChatServerToolChoice): string;
+//# sourceMappingURL=chatservertoolchoice.d.ts.map

package/esm/models/chatservertoolchoice.js

@@ -0,0 +1,13 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ * @generated-id: 19d195c10298
+ */
+import * as z from "zod/v4";
+/** @internal */
+export const ChatServerToolChoice$outboundSchema = z.object({
+    type: z.string(),
+});
+export function chatServerToolChoiceToJSON(chatServerToolChoice) {
+    return JSON.stringify(ChatServerToolChoice$outboundSchema.parse(chatServerToolChoice));
+}
+//# sourceMappingURL=chatservertoolchoice.js.map

package/esm/models/chattoolchoice.d.ts

@@ -1,6 +1,7 @@
 import * as z from "zod/v4";
 import { ClosedEnum } from "../types/enums.js";
 import { ChatNamedToolChoice, ChatNamedToolChoice$Outbound } from "./chatnamedtoolchoice.js";
+import { ChatServerToolChoice, ChatServerToolChoice$Outbound } from "./chatservertoolchoice.js";
 export declare const ChatToolChoiceRequired: {
     readonly Required: "required";
 };
@@ -16,7 +17,7 @@ export type ChatToolChoiceNone = ClosedEnum<typeof ChatToolChoiceNone>;
 /**
  * Tool choice configuration
  */
-export type ChatToolChoice = ChatNamedToolChoice | ChatToolChoiceNone | ChatToolChoiceAuto | ChatToolChoiceRequired;
+export type ChatToolChoice = ChatNamedToolChoice | ChatServerToolChoice | ChatToolChoiceNone | ChatToolChoiceAuto | ChatToolChoiceRequired;
 /** @internal */
 export declare const ChatToolChoiceRequired$outboundSchema: z.ZodEnum<typeof ChatToolChoiceRequired>;
 /** @internal */
@@ -24,7 +25,7 @@ export declare const ChatToolChoiceAuto$outboundSchema: z.ZodEnum<typeof ChatToo
 /** @internal */
 export declare const ChatToolChoiceNone$outboundSchema: z.ZodEnum<typeof ChatToolChoiceNone>;
 /** @internal */
-export type ChatToolChoice$Outbound = ChatNamedToolChoice$Outbound | string | string | string;
+export type ChatToolChoice$Outbound = ChatNamedToolChoice$Outbound | ChatServerToolChoice$Outbound | string | string | string;
 /** @internal */
 export declare const ChatToolChoice$outboundSchema: z.ZodType<ChatToolChoice$Outbound, ChatToolChoice>;
 export declare function chatToolChoiceToJSON(chatToolChoice: ChatToolChoice): string;

package/esm/models/chattoolchoice.js

@@ -4,6 +4,7 @@
  */
 import * as z from "zod/v4";
 import { ChatNamedToolChoice$outboundSchema, } from "./chatnamedtoolchoice.js";
+import { ChatServerToolChoice$outboundSchema, } from "./chatservertoolchoice.js";
 export const ChatToolChoiceRequired = {
     Required: "required",
 };
@@ -22,6 +23,7 @@ export const ChatToolChoiceNone$outboundSchema = z.enum(ChatToolChoiceNone);
 /** @internal */
 export const ChatToolChoice$outboundSchema = z.union([
     ChatNamedToolChoice$outboundSchema,
+    ChatServerToolChoice$outboundSchema,
     ChatToolChoiceNone$outboundSchema,
     ChatToolChoiceAuto$outboundSchema,
     ChatToolChoiceRequired$outboundSchema,

package/esm/models/chatwebsearchshorthand.d.ts

@@ -16,28 +16,32 @@ export type ChatWebSearchShorthandType = OpenEnum<typeof ChatWebSearchShorthandT
  */
 export type ChatWebSearchShorthand = {
     /**
-     * Limit search results to these domains. Supported by Exa, Firecrawl, Parallel, and most native providers (Anthropic, OpenAI, xAI). Not supported with Perplexity. Cannot be used with excluded_domains.
+     * Limit search results to these domains. Supported by Exa, Firecrawl, Parallel, Perplexity, and most native providers (Anthropic, OpenAI, xAI). Cannot be used with excluded_domains.
      */
     allowedDomains?: Array<string> | undefined;
     /**
-     * Which search engine to use. "auto" (default) uses native if the provider supports it, otherwise Exa. "native" forces the provider's built-in search. "exa" forces the Exa search API. "firecrawl" uses Firecrawl (requires BYOK). "parallel" uses the Parallel search API.
+     * Which search engine to use. "auto" (default) uses native if the provider supports it, otherwise Exa. "native" forces the provider's built-in search. "exa" forces the Exa search API. "firecrawl" uses Firecrawl (requires BYOK). "parallel" uses the Parallel search API. "perplexity" uses the Perplexity Search API (raw ranked results).
      */
     engine?: WebSearchEngineEnum | undefined;
     /**
-     * Exclude search results from these domains. Supported by Exa, Firecrawl, Parallel, Anthropic, and xAI. Not supported with OpenAI (silently ignored) or Perplexity. Cannot be used with allowed_domains.
+     * Exclude search results from these domains. Supported by Exa, Firecrawl, Parallel, Perplexity, Anthropic, and xAI. Not supported with OpenAI (silently ignored). Cannot be used with allowed_domains.
      */
     excludedDomains?: Array<string> | undefined;
     /**
-     * Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, and Parallel engines; ignored with native provider search.
+     * Exact maximum number of characters of content per search result. Applies to the Exa, Parallel, and Perplexity engines; ignored with native provider search and Firecrawl. For Exa, caps highlight content per result. For Parallel, caps excerpt content per result (default 1,500 when omitted). For Perplexity, maps to the native `max_tokens_per_page` parameter (converted from characters to tokens) and trims the response to the exact character cap. When both `max_characters` and `search_context_size` are set, `max_characters` takes precedence. When omitted, falls back to `search_context_size` mapping (Exa) or engine defaults (Parallel, Perplexity).
+     */
+    maxCharacters?: number | undefined;
+    /**
+     * Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, Parallel, and Perplexity engines; ignored with native provider search.
      */
     maxResults?: number | undefined;
     /**
-     * Maximum total number of search results across all search calls in a single request. Once this limit is reached, the tool will stop returning new results. Useful for controlling cost and context size in agentic loops.
+     * Maximum total number of search results across all search calls in a single request. Once this limit is reached, the tool will stop returning new results. Useful for controlling cost and context size in agentic loops. Defaults to 50 when not specified.
      */
     maxTotalResults?: number | undefined;
     parameters?: WebSearchConfig | undefined;
     /**
-     * How much context to retrieve per result. Defaults to medium (15000 chars). Applies to Exa and Parallel engines; ignored with native provider search and Firecrawl.
+     * How much context to retrieve per result. Applies to Exa, Parallel, and Perplexity engines; ignored with native provider search and Firecrawl. For Exa, pins a fixed per-result character cap (low=5,000, medium=15,000, high=30,000); when omitted, Exa picks an adaptive size per query and document (typically ~2,000–4,000 characters per result). For Parallel, controls the total characters across all results; when omitted, Parallel uses its own default size. For Perplexity, maps directly to the Search API's native search_context_size parameter. Overridden by `max_characters` when both are set.
      */
     searchContextSize?: SearchQualityLevel | undefined;
     type: ChatWebSearchShorthandType;
@@ -53,6 +57,7 @@ export type ChatWebSearchShorthand$Outbound = {
     allowed_domains?: Array<string> | undefined;
     engine?: string | undefined;
     excluded_domains?: Array<string> | undefined;
+    max_characters?: number | undefined;
     max_results?: number | undefined;
     max_total_results?: number | undefined;
     parameters?: WebSearchConfig$Outbound | undefined;

package/esm/models/chatwebsearchshorthand.js

@@ -22,6 +22,7 @@ export const ChatWebSearchShorthand$outboundSchema = z.object({
     allowedDomains: z.array(z.string()).optional(),
     engine: WebSearchEngineEnum$outboundSchema.optional(),
     excludedDomains: z.array(z.string()).optional(),
+    maxCharacters: z.int().optional(),
     maxResults: z.int().optional(),
     maxTotalResults: z.int().optional(),
     parameters: WebSearchConfig$outboundSchema.optional(),
@@ -32,6 +33,7 @@ export const ChatWebSearchShorthand$outboundSchema = z.object({
     return remap$(v, {
         allowedDomains: "allowed_domains",
         excludedDomains: "excluded_domains",
+        maxCharacters: "max_characters",
         maxResults: "max_results",
         maxTotalResults: "max_total_results",
         searchContextSize: "search_context_size",

package/esm/models/compactionitem.d.ts

@@ -0,0 +1,26 @@
+import * as z from "zod/v4";
+import { ClosedEnum } from "../types/enums.js";
+export declare const CompactionItemType: {
+    readonly Compaction: "compaction";
+};
+export type CompactionItemType = ClosedEnum<typeof CompactionItemType>;
+/**
+ * A context compaction marker with encrypted summary
+ */
+export type CompactionItem = {
+    encryptedContent: string;
+    id?: string | null | undefined;
+    type: CompactionItemType;
+};
+/** @internal */
+export declare const CompactionItemType$outboundSchema: z.ZodEnum<typeof CompactionItemType>;
+/** @internal */
+export type CompactionItem$Outbound = {
+    encrypted_content: string;
+    id?: string | null | undefined;
+    type: string;
+};
+/** @internal */
+export declare const CompactionItem$outboundSchema: z.ZodType<CompactionItem$Outbound, CompactionItem>;
+export declare function compactionItemToJSON(compactionItem: CompactionItem): string;
+//# sourceMappingURL=compactionitem.d.ts.map

package/esm/models/compactionitem.js

@@ -0,0 +1,25 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ * @generated-id: 0c9fb29a5c84
+ */
+import * as z from "zod/v4";
+import { remap as remap$ } from "../lib/primitives.js";
+export const CompactionItemType = {
+    Compaction: "compaction",
+};
+/** @internal */
+export const CompactionItemType$outboundSchema = z.enum(CompactionItemType);
+/** @internal */
+export const CompactionItem$outboundSchema = z.object({
+    encryptedContent: z.string(),
+    id: z.nullable(z.string()).optional(),
+    type: CompactionItemType$outboundSchema,
+}).transform((v) => {
+    return remap$(v, {
+        encryptedContent: "encrypted_content",
+    });
+});
+export function compactionItemToJSON(compactionItem) {
+    return JSON.stringify(CompactionItem$outboundSchema.parse(compactionItem));
+}
+//# sourceMappingURL=compactionitem.js.map

package/esm/models/containerautoenvironment.d.ts

@@ -0,0 +1,15 @@
+import * as z from "zod/v4";
+/**
+ * An OpenRouter-managed, auto-provisioned ephemeral container.
+ */
+export type ContainerAutoEnvironment = {
+    type: "container_auto";
+};
+/** @internal */
+export type ContainerAutoEnvironment$Outbound = {
+    type: "container_auto";
+};
+/** @internal */
+export declare const ContainerAutoEnvironment$outboundSchema: z.ZodType<ContainerAutoEnvironment$Outbound, ContainerAutoEnvironment>;
+export declare function containerAutoEnvironmentToJSON(containerAutoEnvironment: ContainerAutoEnvironment): string;
+//# sourceMappingURL=containerautoenvironment.d.ts.map

package/esm/models/containerautoenvironment.js

@@ -0,0 +1,13 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ * @generated-id: 49526688017e
+ */
+import * as z from "zod/v4";
+/** @internal */
+export const ContainerAutoEnvironment$outboundSchema = z.object({
+    type: z.literal("container_auto"),
+});
+export function containerAutoEnvironmentToJSON(containerAutoEnvironment) {
+    return JSON.stringify(ContainerAutoEnvironment$outboundSchema.parse(containerAutoEnvironment));
+}
+//# sourceMappingURL=containerautoenvironment.js.map

package/esm/models/containerreferenceenvironment.d.ts

@@ -0,0 +1,20 @@
+import * as z from "zod/v4";
+/**
+ * Reference to a previously created container to reuse.
+ */
+export type ContainerReferenceEnvironment = {
+    /**
+     * Identifier of an existing container to reuse (max 20 characters).
+     */
+    containerId: string;
+    type: "container_reference";
+};
+/** @internal */
+export type ContainerReferenceEnvironment$Outbound = {
+    container_id: string;
+    type: "container_reference";
+};
+/** @internal */
+export declare const ContainerReferenceEnvironment$outboundSchema: z.ZodType<ContainerReferenceEnvironment$Outbound, ContainerReferenceEnvironment>;
+export declare function containerReferenceEnvironmentToJSON(containerReferenceEnvironment: ContainerReferenceEnvironment): string;
+//# sourceMappingURL=containerreferenceenvironment.d.ts.map

package/esm/models/containerreferenceenvironment.js

@@ -0,0 +1,19 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ * @generated-id: 711d3c8b4418
+ */
+import * as z from "zod/v4";
+import { remap as remap$ } from "../lib/primitives.js";
+/** @internal */
+export const ContainerReferenceEnvironment$outboundSchema = z.object({
+    containerId: z.string(),
+    type: z.literal("container_reference"),
+}).transform((v) => {
+    return remap$(v, {
+        containerId: "container_id",
+    });
+});
+export function containerReferenceEnvironmentToJSON(containerReferenceEnvironment) {
+    return JSON.stringify(ContainerReferenceEnvironment$outboundSchema.parse(containerReferenceEnvironment));
+}
+//# sourceMappingURL=containerreferenceenvironment.js.map

package/esm/models/contentfilterbuiltinentry.d.ts

@@ -3,6 +3,7 @@ import { Result as SafeParseResult } from "../types/fp.js";
 import { ContentFilterBuiltinAction } from "./contentfilterbuiltinaction.js";
 import { ContentFilterBuiltinSlug } from "./contentfilterbuiltinslug.js";
 import { SDKValidationError } from "./errors/sdkvalidationerror.js";
+import { PromptInjectionScanScope } from "./promptinjectionscanscope.js";
 /**
  * A builtin content filter entry. Builtin filters include PII detectors and the regex-based prompt injection detector.
  */
@@ -12,9 +13,13 @@ export type ContentFilterBuiltinEntry = {
      */
     action: ContentFilterBuiltinAction;
     /**
-     * Optional label used in redaction placeholders (e.g. "[PROMPT_INJECTION]")
+     * Read-only, system-assigned redaction placeholder derived from the slug (e.g. "[EMAIL]", "[PHONE]"). Not settable by the caller.
      */
     label?: string | undefined;
+    /**
+     * Which message roles to scan for prompt injection. Only applies to the regex-prompt-injection builtin. Defaults to all_messages.
+     */
+    scanScope?: PromptInjectionScanScope | undefined;
     /**
      * The builtin filter identifier
      */
@@ -22,14 +27,5 @@ export type ContentFilterBuiltinEntry = {
 };
 /** @internal */
 export declare const ContentFilterBuiltinEntry$inboundSchema: z.ZodType<ContentFilterBuiltinEntry, unknown>;
-/** @internal */
-export type ContentFilterBuiltinEntry$Outbound = {
-    action: string;
-    label?: string | undefined;
-    slug: string;
-};
-/** @internal */
-export declare const ContentFilterBuiltinEntry$outboundSchema: z.ZodType<ContentFilterBuiltinEntry$Outbound, ContentFilterBuiltinEntry>;
-export declare function contentFilterBuiltinEntryToJSON(contentFilterBuiltinEntry: ContentFilterBuiltinEntry): string;
 export declare function contentFilterBuiltinEntryFromJSON(jsonString: string): SafeParseResult<ContentFilterBuiltinEntry, SDKValidationError>;
 //# sourceMappingURL=contentfilterbuiltinentry.d.ts.map