Add support for LLM routing using developer preferences

packages/types/src/global-settings.ts

@@ -166,6 +166,7 @@ export const SECRET_STATE_KEYS = [
 	"xaiApiKey",
 	"groqApiKey",
 	"chutesApiKey",
+	"archgwApiKey",
 	"litellmApiKey",
 	"codeIndexOpenAiKey",
 	"codeIndexQdrantApiKey",

packages/types/src/provider-settings.ts

@@ -24,6 +24,7 @@ export const providerNames = [
 	"mistral",
 	"moonshot",
 	"deepseek",
+	"archgw",
 	"unbound",
 	"requesty",
 	"human-relay",
@@ -195,6 +196,14 @@ const moonshotSchema = apiModelIdProviderModelSchema.extend({
 	moonshotApiKey: z.string().optional(),
 })
 
+const archgwSchema = apiModelIdProviderModelSchema.extend({
+	archgwBaseUrl: z.string().optional(),
+	archgwApiKey: z.string().optional(),
+	archgwModelId: z.string().optional(),
+	archgwUsePreferences: z.boolean().optional(),
+	archgwPreferenceConfig: z.string().optional(),
+})
+
 const unboundSchema = baseProviderSettingsSchema.extend({
 	unboundApiKey: z.string().optional(),
 	unboundModelId: z.string().optional(),
@@ -250,6 +259,7 @@ export const providerSettingsSchemaDiscriminated = z.discriminatedUnion("apiProv
 	mistralSchema.merge(z.object({ apiProvider: z.literal("mistral") })),
 	deepSeekSchema.merge(z.object({ apiProvider: z.literal("deepseek") })),
 	moonshotSchema.merge(z.object({ apiProvider: z.literal("moonshot") })),
+	archgwSchema.merge(z.object({ apiProvider: z.literal("archgw") })),
 	unboundSchema.merge(z.object({ apiProvider: z.literal("unbound") })),
 	requestySchema.merge(z.object({ apiProvider: z.literal("requesty") })),
 	humanRelaySchema.merge(z.object({ apiProvider: z.literal("human-relay") })),
@@ -279,6 +289,7 @@ export const providerSettingsSchema = z.object({
 	...mistralSchema.shape,
 	...deepSeekSchema.shape,
 	...moonshotSchema.shape,
+	...archgwSchema.shape,
 	...unboundSchema.shape,
 	...requestySchema.shape,
 	...humanRelaySchema.shape,
@@ -304,6 +315,7 @@ export const MODEL_ID_KEYS: Partial<keyof ProviderSettings>[] = [
 	"unboundModelId",
 	"requestyModelId",
 	"litellmModelId",
+	"archgwModelId",
 ]
 
 export const getModelId = (settings: ProviderSettings): string | undefined => {

packages/types/src/providers/archgw.ts

@@ -0,0 +1,13 @@
+import type { ModelInfo } from "../model.js"
+
+export const archgwDefaultModelId = "openai/gpt-4.1"
+
+export const archgwDefaultModelInfo: ModelInfo = {
+	maxTokens: 32_768,
+	contextWindow: 1_047_576,
+	supportsImages: true,
+	supportsPromptCache: true,
+	inputPrice: 2,
+	outputPrice: 8,
+	cacheReadsPrice: 0.5,
+}

packages/types/src/providers/index.ts

@@ -18,3 +18,4 @@ export * from "./unbound.js"
 export * from "./vertex.js"
 export * from "./vscode-llm.js"
 export * from "./xai.js"
+export * from "./archgw.js"

scripts/update-contributors.js

@@ -183,14 +183,14 @@ async function readReadme() {
  * @param {Array} contributors Array of contributor objects from GitHub API
  * @returns {string} HTML for contributors section
  */
-const EXCLUDED_LOGIN_SUBSTRINGS = ['[bot]', 'R00-B0T'];
-const EXCLUDED_LOGIN_EXACTS = ['cursor', 'roomote'];
+const EXCLUDED_LOGIN_SUBSTRINGS = ["[bot]", "R00-B0T"]
+const EXCLUDED_LOGIN_EXACTS = ["cursor", "roomote"]
 
 function formatContributorsSection(contributors) {
 	// Filter out GitHub Actions bot, cursor, and roomote
-	const filteredContributors = contributors.filter((c) =>
-		!EXCLUDED_LOGIN_SUBSTRINGS.some(sub => c.login.includes(sub)) &&
-		!EXCLUDED_LOGIN_EXACTS.includes(c.login)
+	const filteredContributors = contributors.filter(
+		(c) =>
+			!EXCLUDED_LOGIN_SUBSTRINGS.some((sub) => c.login.includes(sub)) && !EXCLUDED_LOGIN_EXACTS.includes(c.login),
 	)
 
 	// Start building with Markdown table format

src/api/index.ts

@@ -28,6 +28,7 @@ import {
 	GroqHandler,
 	ChutesHandler,
 	LiteLLMHandler,
+	ArchGwHandler,
 	ClaudeCodeHandler,
 } from "./providers"
 
@@ -92,6 +93,8 @@ export function buildApiHandler(configuration: ProviderSettings): ApiHandler {
 			return new DeepSeekHandler(options)
 		case "moonshot":
 			return new MoonshotHandler(options)
+		case "archgw":
+			return new ArchGwHandler(options)
 		case "vscode-lm":
 			return new VsCodeLmHandler(options)
 		case "mistral":

src/api/providers/__tests__/archgw.spec.ts

@@ -0,0 +1,214 @@
+// npx vitest run src/api/providers/__tests__/archgw.spec.ts
+
+import { Anthropic } from "@anthropic-ai/sdk"
+import { ArchGwHandler } from "../archgw"
+import { ApiHandlerOptions } from "../../../shared/api"
+
+const mockCreate = vitest.fn()
+
+vitest.mock("openai", () => {
+	return {
+		__esModule: true,
+		default: vitest.fn().mockImplementation(() => ({
+			chat: {
+				completions: {
+					create: mockCreate.mockImplementation((options) => {
+						if (!options.stream) {
+							return Promise.resolve({
+								id: "test-completion",
+								choices: [
+									{
+										message: { role: "assistant", content: "ArchGW response" },
+										finish_reason: "stop",
+										index: 0,
+									},
+								],
+								usage: {
+									prompt_tokens: 10,
+									completion_tokens: 5,
+									total_tokens: 15,
+								},
+							})
+						}
+
+						// Streaming: return an object with withResponse() method
+						return {
+							withResponse: () => ({
+								data: {
+									[Symbol.asyncIterator]: async function* () {
+										yield {
+											choices: [
+												{
+													delta: { content: "ArchGW stream" },
+													index: 0,
+												},
+											],
+											usage: null,
+										}
+										yield {
+											choices: [
+												{
+													delta: {},
+													index: 0,
+												},
+											],
+											usage: {
+												prompt_tokens: 10,
+												completion_tokens: 5,
+												total_tokens: 15,
+											},
+										}
+									},
+								},
+							}),
+						}
+					}),
+				},
+			},
+		})),
+	}
+})
+
+describe("ArchGwHandler", () => {
+	let handler: ArchGwHandler
+	let mockOptions: ApiHandlerOptions
+
+	beforeEach(() => {
+		mockOptions = {
+			archgwModelId: "arch-model",
+			archgwBaseUrl: "http://localhost:12000/v1",
+			archgwApiKey: "test-key",
+			archgwPreferenceConfig: "test-pref",
+		}
+		handler = new ArchGwHandler(mockOptions)
+		mockCreate.mockClear()
+	})
+
+	describe("constructor", () => {
+		it("should initialize with provided options", () => {
+			expect(handler).toBeInstanceOf(ArchGwHandler)
+			expect(handler.preferenceConfig).toBe("test-pref")
+		})
+
+		it("should use default base URL if not provided", () => {
+			const handlerWithoutUrl = new ArchGwHandler({
+				archgwModelId: "arch-model",
+				archgwApiKey: "test-key",
+			})
+			expect(handlerWithoutUrl).toBeInstanceOf(ArchGwHandler)
+		})
+
+		it("should not set preferenceConfig if not provided", () => {
+			const handlerNoPref = new ArchGwHandler({
+				archgwModelId: "arch-model",
+				archgwApiKey: "test-key",
+			})
+			expect(handlerNoPref.preferenceConfig).toBeUndefined()
+		})
+
+		it("should set preferenceConfig to empty string if provided as empty", () => {
+			const handlerEmptyPref = new ArchGwHandler({
+				archgwModelId: "arch-model",
+				archgwApiKey: "test-key",
+				archgwPreferenceConfig: "",
+			})
+			expect(handlerEmptyPref.preferenceConfig).toBe("")
+		})
+	})
+
+	describe("createMessage", () => {
+		const systemPrompt = "You are a helpful assistant."
+		const messages: Anthropic.Messages.MessageParam[] = [
+			{
+				role: "user",
+				content: "Hello!",
+			},
+		]
+
+		it("should handle streaming responses", async () => {
+			const stream = handler.createMessage(systemPrompt, messages)
+			const chunks: any[] = []
+			for await (const chunk of stream) {
+				chunks.push(chunk)
+			}
+
+			expect(chunks.length).toBeGreaterThan(0)
+			const textChunks = chunks.filter((chunk) => chunk.type === "text")
+			expect(textChunks).toHaveLength(1)
+			expect(textChunks[0].text).toBe("ArchGW stream")
+		})
+
+		it("should handle API errors", async () => {
+			mockCreate.mockImplementationOnce(() => ({
+				withResponse: () => {
+					throw new Error("API Error")
+				},
+			}))
+
+			const stream = handler.createMessage(systemPrompt, messages)
+
+			await expect(async () => {
+				for await (const _chunk of stream) {
+					// Should not reach here
+				}
+			}).rejects.toThrow("archgw streaming error: API Error")
+		})
+	})
+
+	describe("completePrompt", () => {
+		it("should complete prompt successfully", async () => {
+			const result = await handler.completePrompt("Test prompt")
+			expect(result).toBe("ArchGW response")
+			expect(mockCreate).toHaveBeenCalledWith({
+				model: mockOptions.archgwModelId,
+				messages: [{ role: "user", content: "Test prompt" }],
+				temperature: 0,
+				max_tokens: 32768,
+				metadata: { archgw_preference_config: "test-pref" },
+			})
+		})
+
+		it("should not include archgw_preference_config in metadata if preferenceConfig is not set", async () => {
+			const handlerNoPref = new ArchGwHandler({
+				archgwModelId: "arch-model",
+				archgwApiKey: "test-key",
+			})
+			mockCreate.mockClear()
+			await handlerNoPref.completePrompt("Test prompt")
+			const call = mockCreate.mock.calls[0][0]
+			expect(call.metadata).toBeUndefined()
+		})
+
+		it("should include archgw_preference_config in metadata if preferenceConfig is empty string", async () => {
+			const preferenceConfig = `- name: code generation
+  model: gpt-4o-mini
+  usage: generating new code snippets
+- name: code understanding
+  model: gpt-4.1
+  usage: understand and explain existing code snippets
+`
+			const handlerEmptyPref = new ArchGwHandler({
+				archgwModelId: "arch-modael",
+				archgwApiKey: "test-key",
+				archgwPreferenceConfig: preferenceConfig,
+			})
+			mockCreate.mockClear()
+			await handlerEmptyPref.completePrompt("Test prompt")
+			const call = mockCreate.mock.calls[0][0]
+			expect(call.metadata).toEqual({ archgw_preference_config: preferenceConfig })
+		})
+
+		it("should handle API errors", async () => {
+			mockCreate.mockRejectedValueOnce(new Error("API Error"))
+			await expect(handler.completePrompt("Test prompt")).rejects.toThrow("archgw completion error: API Error")
+		})
+
+		it("should handle empty response", async () => {
+			mockCreate.mockResolvedValueOnce({
+				choices: [{ message: { content: "" } }],
+			})
+			const result = await handler.completePrompt("Test prompt")
+			expect(result).toBe("")
+		})
+	})
+})