@hebo-ai/gateway 0.10.7 → 0.11.0

package/dist/providers/fireworks/canonical.js

@@ -0,0 +1,27 @@
+import {} from "@ai-sdk/fireworks";
+import { withCanonicalIds } from "../registry";
+const MAPPING = {
+    "meta/llama-3.1-8b": "accounts/fireworks/models/llama-v3p1-8b-instruct",
+    "meta/llama-3.1-405b": "accounts/fireworks/models/llama-v3p1-405b-instruct",
+    "meta/llama-3.2-3b": "accounts/fireworks/models/llama-v3p2-3b-instruct",
+    "meta/llama-3.2-11b": "accounts/fireworks/models/llama-v3p2-11b-vision-instruct",
+    "meta/llama-3.3-70b": "accounts/fireworks/models/llama-v3p3-70b-instruct",
+    "openai/gpt-oss-20b": "accounts/fireworks/models/gpt-oss-20b",
+    "openai/gpt-oss-120b": "accounts/fireworks/models/gpt-oss-120b",
+    "minimax/m2.7": "accounts/fireworks/models/minimax-m2p7",
+    "alibaba/qwen3-235b": "accounts/fireworks/models/qwen3-235b-a22b",
+    "alibaba/qwen3-32b": "accounts/fireworks/models/qwen3-32b",
+    "alibaba/qwen3.5-397b": "accounts/fireworks/models/qwen3p5-397b-a17b",
+    "alibaba/qwen3.5-35b": "accounts/fireworks/models/qwen3p5-35b-a3b",
+    "alibaba/qwen3.5-27b": "accounts/fireworks/models/qwen3p5-27b",
+    "alibaba/qwen3.5-9b": "accounts/fireworks/models/qwen3p5-9b",
+    "deepseek/deepseek-v3.2": "accounts/fireworks/models/deepseek-v3p2",
+    "moonshot/kimi-k2.5": "accounts/fireworks/models/kimi-k2p5",
+    "moonshot/kimi-k2.6": "accounts/fireworks/models/kimi-k2p6",
+    "zhipu/glm-5": "accounts/fireworks/models/glm-5",
+    "zhipu/glm-5.1": "accounts/fireworks/models/glm-5p1",
+};
+export const withCanonicalIdsForFireworks = (provider, extraMapping) => withCanonicalIds(provider, {
+    mapping: { ...MAPPING, ...extraMapping },
+    options: { stripNamespace: false },
+});

package/dist/providers/fireworks/index.d.ts

@@ -0,0 +1,2 @@
+export * from "./canonical";
+export * from "./middleware";

package/dist/providers/fireworks/index.js

@@ -0,0 +1,2 @@
+export * from "./canonical";
+export * from "./middleware";

package/dist/providers/fireworks/middleware.d.ts

@@ -0,0 +1,2 @@
+import type { LanguageModelMiddleware } from "ai";
+export declare const fireworksReasoningMiddleware: LanguageModelMiddleware;

package/dist/providers/fireworks/middleware.js

@@ -0,0 +1,35 @@
+import { modelMiddlewareMatcher } from "../../middleware/matcher";
+import { calculateReasoningBudgetFromEffort } from "../../middleware/utils";
+const FIREWORKS_MAX_OUTPUT_TOKENS = 131072;
+export const fireworksReasoningMiddleware = {
+    specificationVersion: "v3",
+    // oxlint-disable-next-line require-await
+    transformParams: async ({ params }) => {
+        const fireworks = params.providerOptions?.["fireworks"];
+        if (!fireworks || typeof fireworks !== "object")
+            return params;
+        const reasoning = fireworks["reasoning"];
+        if (!reasoning)
+            return params;
+        if (!reasoning.enabled) {
+            fireworks.thinking = { type: "disabled" };
+        }
+        else if (reasoning.max_tokens) {
+            fireworks.thinking = { type: "enabled", budgetTokens: reasoning.max_tokens };
+        }
+        else if (reasoning.effort) {
+            fireworks.thinking = {
+                type: "enabled",
+                budgetTokens: calculateReasoningBudgetFromEffort(reasoning.effort, params.maxOutputTokens ?? FIREWORKS_MAX_OUTPUT_TOKENS, 1024),
+            };
+        }
+        else {
+            fireworks.thinking = { type: "enabled" };
+        }
+        delete fireworks["reasoning"];
+        return params;
+    },
+};
+modelMiddlewareMatcher.useForProvider("fireworks", {
+    language: [fireworksReasoningMiddleware],
+});

package/dist/providers/groq/canonical.js

@@ -4,7 +4,7 @@ const MAPPING = {
     "meta/llama-3.1-8b": "llama-3.1-8b-instant",
     "meta/llama-3.3-70b": "llama-3.3-70b-versatile",
     "meta/llama-4-scout": "meta-llama/llama-4-scout-17b-16e-instruct",
-    "meta/llama-4-maverick": "meta-llama/llama-4-maverick-17b-128e-instruct",
+    "alibaba/qwen3-32b": "qwen/qwen3-32b",
 };
 export const withCanonicalIdsForGroq = (provider, extraMapping) => withCanonicalIds(provider, {
     mapping: { ...MAPPING, ...extraMapping },

package/dist/providers/minimax/canonical.d.ts

@@ -0,0 +1,3 @@
+import type { ProviderV3 } from "@ai-sdk/provider";
+import type { ModelId } from "../../models/types";
+export declare const withCanonicalIdsForMinimax: (provider: ProviderV3, extraMapping?: Record<ModelId, string>) => ProviderV3;

package/dist/providers/minimax/canonical.js

@@ -0,0 +1,9 @@
+import { withCanonicalIds } from "../registry";
+const MAPPING = {
+    "minimax/m2.5": "MiniMax-M2.5",
+    "minimax/m2.7": "MiniMax-M2.7",
+};
+export const withCanonicalIdsForMinimax = (provider, extraMapping) => withCanonicalIds(provider, {
+    mapping: { ...MAPPING, ...extraMapping },
+    options: { stripNamespace: false },
+});

package/dist/providers/minimax/index.d.ts

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/minimax/index.js

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/moonshot/canonical.d.ts

@@ -0,0 +1,3 @@
+import { type MoonshotAIProvider } from "@ai-sdk/moonshotai";
+import type { ModelId } from "../../models/types";
+export declare const withCanonicalIdsForMoonshot: (provider: MoonshotAIProvider, extraMapping?: Partial<Record<ModelId, string>>) => import("@ai-sdk/provider").ProviderV3;

package/dist/providers/moonshot/canonical.js

@@ -0,0 +1,6 @@
+import {} from "@ai-sdk/moonshotai";
+import { withCanonicalIds } from "../registry";
+export const withCanonicalIdsForMoonshot = (provider, extraMapping) => withCanonicalIds(provider, {
+    mapping: extraMapping,
+    options: { stripNamespace: true },
+});

package/dist/providers/moonshot/index.d.ts

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/moonshot/index.js

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/togetherai/canonical.d.ts

@@ -0,0 +1,3 @@
+import { type TogetherAIProvider } from "@ai-sdk/togetherai";
+import type { ModelId } from "../../models/types";
+export declare const withCanonicalIdsForTogetherAI: (provider: TogetherAIProvider, extraMapping?: Record<ModelId, string>) => import("@ai-sdk/provider").ProviderV3;

package/dist/providers/togetherai/canonical.js

@@ -0,0 +1,24 @@
+import {} from "@ai-sdk/togetherai";
+import { withCanonicalIds } from "../registry";
+const MAPPING = {
+    "google/gemma-4-31b": "google/gemma-4-31B-it",
+    "meta/llama-3.1-8b": "meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo",
+    "meta/llama-3.1-70b": "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",
+    "meta/llama-3.1-405b": "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo",
+    "meta/llama-3.2-3b": "meta-llama/Llama-3.2-3B-Instruct-Turbo",
+    "meta/llama-3.3-70b": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
+    "meta/llama-4-maverick": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8",
+    "minimax/m2.5": "MiniMax/MiniMax-M2.5",
+    "minimax/m2.7": "MiniMax/MiniMax-M2.7",
+    "alibaba/qwen3-235b": "Qwen/Qwen3-235B-A22B",
+    "alibaba/qwen3-32b": "Qwen/Qwen3-32B",
+    "alibaba/qwen3.5-397b": "Qwen/Qwen3.5-397B-A17B",
+    "alibaba/qwen3.5-9b": "Qwen/Qwen3.5-9B",
+    "moonshot/kimi-k2.5": "moonshotai/Kimi-K2.5",
+    "zhipu/glm-5": "zai-org/GLM-5",
+    "zhipu/glm-5.1": "zai-org/GLM-5.1",
+};
+export const withCanonicalIdsForTogetherAI = (provider, extraMapping) => withCanonicalIds(provider, {
+    mapping: { ...MAPPING, ...extraMapping },
+    options: { stripNamespace: false },
+});

package/dist/providers/togetherai/index.d.ts

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/togetherai/index.js

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/types.d.ts

@@ -1,5 +1,5 @@
 import { type ProviderV3 } from "@ai-sdk/provider";
-export declare const CANONICAL_PROVIDER_IDS: readonly ["anthropic", "azure", "bedrock", "cohere", "groq", "openai", "vertex", "voyage"];
+export declare const CANONICAL_PROVIDER_IDS: readonly ["alibaba", "anthropic", "azure", "bedrock", "chutes", "cohere", "deepinfra", "deepseek", "fireworks", "groq", "minimax", "moonshot", "openai", "togetherai", "vertex", "voyage", "xai", "zai"];
 export type CanonicalProviderId = (typeof CANONICAL_PROVIDER_IDS)[number];
 export type ProviderId = CanonicalProviderId | (string & {});
 export type ProviderRegistry = {

package/dist/providers/types.js

@@ -1,11 +1,21 @@
 import {} from "@ai-sdk/provider";
 export const CANONICAL_PROVIDER_IDS = [
+    "alibaba",
     "anthropic",
     "azure",
     "bedrock",
+    "chutes",
     "cohere",
+    "deepinfra",
+    "deepseek",
+    "fireworks",
     "groq",
+    "minimax",
+    "moonshot",
     "openai",
+    "togetherai",
     "vertex",
     "voyage",
+    "xai",
+    "zai",
 ];

package/dist/providers/vertex/canonical.js

@@ -1,6 +1,10 @@
 import { withCanonicalIds } from "../registry";
+const MAPPING = {
+    "alibaba/qwen3-235b": "qwen3-235b-a22b-instruct-2507-maas",
+    "deepseek/deepseek-v3.2": "deepseek-v3.2-maas",
+};
 export const withCanonicalIdsForVertex = (provider, extraMapping) => withCanonicalIds(provider, {
-    mapping: extraMapping,
+    mapping: { ...MAPPING, ...extraMapping },
     options: {
         stripNamespace: true,
         normalizeDelimiters: ["anthropic"],

package/dist/providers/xai/canonical.d.ts

@@ -0,0 +1,3 @@
+import { type XaiProvider } from "@ai-sdk/xai";
+import type { ModelId } from "../../models/types";
+export declare const withCanonicalIdsForXai: (provider: XaiProvider, extraMapping?: Partial<Record<ModelId, string>>) => import("@ai-sdk/provider").ProviderV3;

package/dist/providers/xai/canonical.js

@@ -0,0 +1,12 @@
+import {} from "@ai-sdk/xai";
+import { withCanonicalIds } from "../registry";
+const MAPPING = {
+    "xai/grok-4.1-fast": "grok-4-1-fast-non-reasoning",
+    "xai/grok-4.2": "grok-4.20-0309-non-reasoning",
+    "xai/grok-4.2-reasoning": "grok-4.20-0309-reasoning",
+    "xai/grok-4.2-multi-agent": "grok-4.20-multi-agent-0309",
+};
+export const withCanonicalIdsForXai = (provider, extraMapping) => withCanonicalIds(provider, {
+    mapping: { ...MAPPING, ...extraMapping },
+    options: { stripNamespace: true, normalizeDelimiters: true },
+});

package/dist/providers/xai/index.d.ts

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/xai/index.js

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/zai/canonical.d.ts

@@ -0,0 +1,3 @@
+import type { ZhipuProvider } from "zhipu-ai-provider";
+import type { ModelId } from "../../models/types";
+export declare const withCanonicalIdsForZai: (provider: ZhipuProvider, extraMapping?: Partial<Record<ModelId, string>>) => import("@ai-sdk/provider").ProviderV3;

package/dist/providers/zai/canonical.js

@@ -0,0 +1,10 @@
+import { withCanonicalIds } from "../registry";
+const MAPPING = {
+    "zhipu/glm-5": "glm-5-20260211",
+    "zhipu/glm-5-turbo": "glm-5-turbo-20260315",
+    "zhipu/glm-5.1": "glm-5.1-20260406",
+};
+export const withCanonicalIdsForZai = (provider, extraMapping) => withCanonicalIds(provider, {
+    mapping: { ...MAPPING, ...extraMapping },
+    options: { stripNamespace: true },
+});

package/dist/providers/zai/index.d.ts

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/providers/zai/index.js

@@ -0,0 +1 @@
+export * from "./canonical";

package/dist/types.d.ts

@@ -209,27 +209,41 @@ export type GatewayConfig = {
         };
     };
     /**
-     * Optional timeout for server responses.
-     * Supports a number in milliseconds, or tiered config.
+     * Advanced configuration for timeouts, body limits, and header forwarding.
      */
-    timeouts?: GatewayTimeout;
-    /**
-     * Maximum *decompressed* request body size in bytes for gzip/deflate-encoded requests.
-     * Plain (uncompressed) request body size limits should be configured at the
-     * framework or server level (e.g. Hono `bodyLimit` middleware, Bun `maxRequestBodySize`).
-     * Set to `0` to disable the decompressed size limit.
-     * Defaults to 10 MB (10,485,760 bytes).
-     */
-    maxBodySize?: number;
+    advanced?: {
+        /**
+         * Optional timeout for server responses.
+         * Supports a number in milliseconds, or tiered config.
+         */
+        timeouts?: GatewayTimeout;
+        /**
+         * Maximum *decompressed* request body size in bytes for gzip/deflate-encoded requests.
+         * Plain (uncompressed) request body size limits should be configured at the
+         * framework or server level (e.g. Hono `bodyLimit` middleware, Bun `maxRequestBodySize`).
+         * Set to `0` to disable the decompressed size limit.
+         * Defaults to 10 MB (10,485,760 bytes).
+         */
+        maxBodySize?: number;
+        /**
+         * Additional headers to forward to upstream providers,
+         * merged with the built-in allowlist at startup.
+         * Header names are matched case-insensitively.
+         */
+        forwardHeaders?: string[];
+    };
 };
 export declare const kParsed: unique symbol;
-export type GatewayConfigParsed = Omit<GatewayConfig, "storage" | "timeouts"> & {
+export type GatewayConfigParsed = Omit<GatewayConfig, "storage" | "advanced"> & {
     storage: ConversationStorage;
-    timeouts: {
-        normal?: number;
-        flex?: number;
+    advanced: {
+        timeouts: {
+            normal?: number;
+            flex?: number;
+        };
+        maxBodySize: number;
+        forwardHeaders: string[];
     };
-    maxBodySize: number;
     [kParsed]: true;
 };
 export interface Endpoint {

package/dist/utils/index.d.ts

@@ -0,0 +1,8 @@
+export * from "./body";
+export * from "./env";
+export * from "./headers";
+export * from "./preset";
+export * from "./request";
+export * from "./response";
+export * from "./stream";
+export * from "./url";

package/dist/utils/index.js

@@ -0,0 +1,8 @@
+export * from "./body";
+export * from "./env";
+export * from "./headers";
+export * from "./preset";
+export * from "./request";
+export * from "./response";
+export * from "./stream";
+export * from "./url";

package/dist/utils/request.d.ts

@@ -1,2 +1,3 @@
+export declare const FORWARD_HEADER_ALLOWLIST: readonly ["openai-beta", "openai-organization", "openai-project", "or_app_name", "or_site_url", "x-openrouter-categories", "x-openrouter-title", "x-title", "anthropic-beta", "anthropic-version", "x-claude-code-session-id", "x-amzn-bedrock-guardrailidentifier", "x-amzn-bedrock-guardrailversion", "x-amzn-bedrock-performanceconfig-latency", "x-amzn-bedrock-trace", "x-goog-api-client", "x-vertex-ai-endpoint-id", "x-vertex-ai-llm-request-type", "x-vertex-ai-llm-shared-request-type", "x-kilo-session", "x-kilocode-taskid", "x-kilocode-editorname", "x-kilocode-feature", "x-kilocode-machineid", "x-kilocode-organizationid", "x-kilocode-projectid", "x-kilocode-tester", "agent-session-id", "x-client", "x-client-name", "x-client-type", "x-client-version", "x-platform", "x-platform-version", "x-task-id", "x-stainless-arch", "x-stainless-lang", "x-stainless-os", "x-stainless-package-version", "x-stainless-runtime", "x-stainless-runtime-version"];
 export declare const resolveOrCreateRequestId: (request: Request) => string;
-export declare const prepareForwardHeaders: (request: Request) => Record<string, string>;
+export declare const prepareForwardHeaders: (request: Request, allowlist?: readonly string[]) => Record<string, string>;

package/dist/utils/request.js

@@ -1,32 +1,60 @@
 import pkg from "../../package.json" with { type: "json" };
 import { resolveRequestId } from "./headers";
 const GATEWAY_VERSION = pkg.version;
-const FORWARD_HEADER_ALLOWLIST = [
-    // OpenAI + OpenAI-compatible providers (Azure, Groq, Together, Fireworks, etc.)
+export const FORWARD_HEADER_ALLOWLIST = [
+    // OpenAI
     "openai-beta",
     "openai-organization",
     "openai-project",
     // OpenRouter
+    "or_app_name",
+    "or_site_url",
     "x-openrouter-categories",
     "x-openrouter-title",
     "x-title",
     // Anthropic
     "anthropic-beta",
+    "anthropic-version",
+    "x-claude-code-session-id",
     // AWS Bedrock
     "x-amzn-bedrock-guardrailidentifier",
     "x-amzn-bedrock-guardrailversion",
     "x-amzn-bedrock-performanceconfig-latency",
     "x-amzn-bedrock-trace",
-    // Cohere
-    "x-client-name",
-    // Vertex provisioned throughput / endpoint routing
+    // Google Vertex
+    "x-goog-api-client",
     "x-vertex-ai-endpoint-id",
     "x-vertex-ai-llm-request-type",
     "x-vertex-ai-llm-shared-request-type",
+    // Kilocode
+    "x-kilo-session",
+    "x-kilocode-taskid",
+    "x-kilocode-editorname",
+    "x-kilocode-feature",
+    "x-kilocode-machineid",
+    "x-kilocode-organizationid",
+    "x-kilocode-projectid",
+    "x-kilocode-tester",
+    // Agent / org / project / session ids
+    "agent-session-id",
+    "x-client",
+    "x-client-name",
+    "x-client-type",
+    "x-client-version",
+    "x-platform",
+    "x-platform-version",
+    "x-task-id",
+    // SDK / protocol identification
+    "x-stainless-arch",
+    "x-stainless-lang",
+    "x-stainless-os",
+    "x-stainless-package-version",
+    "x-stainless-runtime",
+    "x-stainless-runtime-version",
 ];
 const createRequestId = () => "req_" + crypto.getRandomValues(new Uint32Array(2)).reduce((s, n) => s + n.toString(36), "");
 export const resolveOrCreateRequestId = (request) => resolveRequestId(request) ?? createRequestId();
-export const prepareForwardHeaders = (request) => {
+export const prepareForwardHeaders = (request, allowlist = FORWARD_HEADER_ALLOWLIST) => {
     const userAgent = request.headers.get("user-agent");
     const appendedUserAgent = userAgent
         ? `${userAgent} @hebo-ai/gateway/${GATEWAY_VERSION}`
@@ -34,7 +62,7 @@ export const prepareForwardHeaders = (request) => {
     const headers = {
         "user-agent": appendedUserAgent,
     };
-    for (const key of FORWARD_HEADER_ALLOWLIST) {
+    for (const key of allowlist) {
         const value = request.headers.get(key);
         if (value !== null)
             headers[key] = value;