Add abliteration.ai provider

.changeset/abliteration-provider.md

@@ -0,0 +1,5 @@
+---
+"roo-cline": patch
+---
+
+Add abliteration.ai as a provider.

apps/cli/src/lib/utils/__tests__/provider.test.ts

@@ -17,6 +17,11 @@ describe("getApiKeyFromEnv", () => {
 		expect(getApiKeyFromEnv("anthropic")).toBe("test-anthropic-key")
 	})
 
+	it("should return API key from environment variable for abliteration.ai", () => {
+		process.env.ABLIT_KEY = "test-abliteration-key"
+		expect(getApiKeyFromEnv("abliteration")).toBe("test-abliteration-key")
+	})
+
 	it("should return API key from environment variable for openrouter", () => {
 		process.env.OPENROUTER_API_KEY = "test-openrouter-key"
 		expect(getApiKeyFromEnv("openrouter")).toBe("test-openrouter-key")

apps/cli/src/lib/utils/provider.ts

@@ -3,6 +3,7 @@ import { RooCodeSettings } from "@roo-code/types"
 import type { SupportedProvider } from "@/types/index.js"
 
 const envVarMap: Record<SupportedProvider, string> = {
+	abliteration: "ABLIT_KEY",
 	anthropic: "ANTHROPIC_API_KEY",
 	"openai-native": "OPENAI_API_KEY",
 	gemini: "GOOGLE_API_KEY",
@@ -28,6 +29,10 @@ export function getProviderSettings(
 	const config: RooCodeSettings = { apiProvider: provider }
 
 	switch (provider) {
+		case "abliteration":
+			if (apiKey) config.abliterationApiKey = apiKey
+			if (model) config.apiModelId = model
+			break
 		case "anthropic":
 			if (apiKey) config.apiKey = apiKey
 			if (model) config.apiModelId = model

apps/cli/src/types/types.ts

@@ -2,6 +2,7 @@ import type { ProviderName, ReasoningEffortExtended } from "@roo-code/types"
 import type { OutputFormat } from "./json-events.js"
 
 export const supportedProviders = [
+	"abliteration",
 	"anthropic",
 	"openai-native",
 	"gemini",

packages/types/src/global-settings.ts

@@ -252,6 +252,7 @@ export type RooCodeSettings = GlobalSettings & ProviderSettings
 export const SECRET_STATE_KEYS = [
 	"apiKey",
 	"openRouterApiKey",
+	"abliterationApiKey",
 	"awsAccessKey",
 	"awsApiKey",
 	"awsSecretKey",

packages/types/src/provider-settings.ts

@@ -3,6 +3,7 @@ import { z } from "zod"
 import { modelInfoSchema, reasoningEffortSettingSchema, verbosityLevelsSchema, serviceTierSchema } from "./model.js"
 import { codebaseIndexProviderSchema } from "./codebase-index.js"
 import {
+	abliterationModels,
 	anthropicModels,
 	basetenModels,
 	bedrockModels,
@@ -110,6 +111,7 @@ export const providerNames = [
 	...internalProviders,
 	...customProviders,
 	...fauxProviders,
+	"abliteration",
 	"anthropic",
 	"bedrock",
 	"baseten",
@@ -209,6 +211,10 @@ const anthropicSchema = apiModelIdProviderModelSchema.extend({
 	anthropicBeta1MContext: z.boolean().optional(), // Enable 'context-1m-2025-08-07' beta for 1M context window.
 })
 
+const abliterationSchema = apiModelIdProviderModelSchema.extend({
+	abliterationApiKey: z.string().optional(),
+})
+
 const openRouterSchema = baseProviderSettingsSchema.extend({
 	openRouterApiKey: z.string().optional(),
 	openRouterModelId: z.string().optional(),
@@ -399,6 +405,7 @@ const defaultSchema = z.object({
 })
 
 export const providerSettingsSchemaDiscriminated = z.discriminatedUnion("apiProvider", [
+	abliterationSchema.merge(z.object({ apiProvider: z.literal("abliteration") })),
 	anthropicSchema.merge(z.object({ apiProvider: z.literal("anthropic") })),
 	openRouterSchema.merge(z.object({ apiProvider: z.literal("openrouter") })),
 	bedrockSchema.merge(z.object({ apiProvider: z.literal("bedrock") })),
@@ -433,6 +440,7 @@ export const providerSettingsSchemaDiscriminated = z.discriminatedUnion("apiProv
 
 export const providerSettingsSchema = z.object({
 	apiProvider: providerNamesWithRetiredSchema.optional(),
+	...abliterationSchema.shape,
 	...anthropicSchema.shape,
 	...openRouterSchema.shape,
 	...bedrockSchema.shape,
@@ -511,6 +519,7 @@ export const isTypicalProvider = (key: unknown): key is TypicalProvider =>
 	isProviderName(key) && !isInternalProvider(key) && !isCustomProvider(key) && !isFauxProvider(key)
 
 export const modelIdKeysByProvider: Record<TypicalProvider, ModelIdKey> = {
+	abliteration: "apiModelId",
 	anthropic: "apiModelId",
 	openrouter: "openRouterModelId",
 	bedrock: "apiModelId",
@@ -576,6 +585,11 @@ export const MODELS_BY_PROVIDER: Record<
 	Exclude<ProviderName, "fake-ai" | "gemini-cli" | "openai">,
 	{ id: ProviderName; label: string; models: string[] }
 > = {
+	abliteration: {
+		id: "abliteration",
+		label: "abliteration.ai",
+		models: Object.keys(abliterationModels),
+	},
 	anthropic: {
 		id: "anthropic",
 		label: "Anthropic",

packages/types/src/providers/abliteration.ts

@@ -0,0 +1,17 @@
+import type { ModelInfo } from "../model.js"
+
+// https://docs.abliteration.ai/models
+export type AbliterationModelId = "abliterated-model"
+
+export const abliterationDefaultModelId: AbliterationModelId = "abliterated-model"
+
+export const abliterationModels = {
+	"abliterated-model": {
+		maxTokens: 8192,
+		contextWindow: 150_000,
+		supportsImages: true,
+		supportsPromptCache: false,
+		description:
+			"Default abliteration.ai model. Supports OpenAI-compatible chat completions, streaming, tool calling, and vision.",
+	},
+} as const satisfies Record<string, ModelInfo>

packages/types/src/providers/index.ts

@@ -1,3 +1,4 @@
+export * from "./abliteration.js"
 export * from "./anthropic.js"
 export * from "./baseten.js"
 export * from "./bedrock.js"
@@ -26,6 +27,7 @@ export * from "./vercel-ai-gateway.js"
 export * from "./zai.js"
 export * from "./minimax.js"
 
+import { abliterationDefaultModelId } from "./abliteration.js"
 import { anthropicDefaultModelId } from "./anthropic.js"
 import { basetenDefaultModelId } from "./baseten.js"
 import { bedrockDefaultModelId } from "./bedrock.js"
@@ -63,6 +65,8 @@ export function getProviderDefaultModelId(
 	options: { isChina?: boolean } = { isChina: false },
 ): string {
 	switch (provider) {
+		case "abliteration":
+			return abliterationDefaultModelId
 		case "openrouter":
 			return openRouterDefaultModelId
 		case "requesty":

src/api/index.ts

@@ -7,6 +7,7 @@ import { ApiStream } from "./transform/stream"
 
 import {
 	AnthropicHandler,
+	AbliterationHandler,
 	AwsBedrockHandler,
 	OpenRouterHandler,
 	PoeHandler,
@@ -119,6 +120,8 @@ export function buildApiHandler(configuration: ProviderSettings): ApiHandler {
 	}
 
 	switch (apiProvider) {
+		case "abliteration":
+			return new AbliterationHandler(options)
 		case "anthropic":
 			return new AnthropicHandler(options)
 		case "openrouter":

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

@@ -0,0 +1,129 @@
+// npx vitest run src/api/providers/__tests__/abliteration.spec.ts
+
+import { Anthropic } from "@anthropic-ai/sdk"
+import OpenAI from "openai"
+
+import { abliterationDefaultModelId, abliterationModels } from "@roo-code/types"
+
+import { AbliterationHandler } from "../abliteration"
+
+const mockCreate = vi.fn()
+
+vi.mock("openai", () => ({
+	default: vi.fn(() => ({
+		chat: {
+			completions: {
+				create: mockCreate,
+			},
+		},
+	})),
+}))
+
+describe("AbliterationHandler", () => {
+	let handler: AbliterationHandler
+
+	beforeEach(() => {
+		vi.clearAllMocks()
+		handler = new AbliterationHandler({ abliterationApiKey: "test-abliteration-api-key" })
+	})
+
+	it("should use the correct abliteration.ai base URL", () => {
+		new AbliterationHandler({ abliterationApiKey: "test-abliteration-api-key" })
+		expect(OpenAI).toHaveBeenCalledWith(expect.objectContaining({ baseURL: "https://api.abliteration.ai/v1" }))
+	})
+
+	it("should use the provided API key", () => {
+		const abliterationApiKey = "test-abliteration-api-key"
+		new AbliterationHandler({ abliterationApiKey })
+		expect(OpenAI).toHaveBeenCalledWith(expect.objectContaining({ apiKey: abliterationApiKey }))
+	})
+
+	it("should throw error when API key is not provided", () => {
+		expect(() => new AbliterationHandler({})).toThrow("API key is required")
+	})
+
+	it("should return default model when no model is specified", () => {
+		const model = handler.getModel()
+		expect(model.id).toBe(abliterationDefaultModelId)
+		expect(model.info).toEqual(abliterationModels[abliterationDefaultModelId])
+	})
+
+	it("should return specified model when valid model is provided", () => {
+		const handlerWithModel = new AbliterationHandler({
+			apiModelId: "abliterated-model",
+			abliterationApiKey: "test-abliteration-api-key",
+		})
+		const model = handlerWithModel.getModel()
+		expect(model.id).toBe("abliterated-model")
+		expect(model.info).toEqual(abliterationModels["abliterated-model"])
+	})
+
+	it("completePrompt method should return text from abliteration.ai API", async () => {
+		const expectedResponse = "This is a test response from abliteration.ai"
+		mockCreate.mockResolvedValueOnce({ choices: [{ message: { content: expectedResponse } }] })
+		const result = await handler.completePrompt("test prompt")
+		expect(result).toBe(expectedResponse)
+	})
+
+	it("createMessage should yield text content from stream", async () => {
+		const testContent = "This is test content from abliteration.ai stream"
+
+		mockCreate.mockImplementationOnce(() => {
+			return {
+				[Symbol.asyncIterator]: () => ({
+					next: vi
+						.fn()
+						.mockResolvedValueOnce({
+							done: false,
+							value: { choices: [{ delta: { content: testContent } }] },
+						})
+						.mockResolvedValueOnce({ done: true }),
+				}),
+			}
+		})
+
+		const stream = handler.createMessage("system prompt", [])
+		const firstChunk = await stream.next()
+
+		expect(firstChunk.done).toBe(false)
+		expect(firstChunk.value).toEqual({ type: "text", text: testContent })
+	})
+
+	it("createMessage should pass correct parameters to abliteration.ai client", async () => {
+		const modelInfo = abliterationModels[abliterationDefaultModelId]
+		const handlerWithModel = new AbliterationHandler({
+			apiModelId: abliterationDefaultModelId,
+			abliterationApiKey: "test-abliteration-api-key",
+		})
+
+		mockCreate.mockImplementationOnce(() => {
+			return {
+				[Symbol.asyncIterator]: () => ({
+					async next() {
+						return { done: true }
+					},
+				}),
+			}
+		})
+
+		const systemPrompt = "Test system prompt for abliteration.ai"
+		const messages: Anthropic.Messages.MessageParam[] = [
+			{ role: "user", content: "Test message for abliteration.ai" },
+		]
+
+		const messageGenerator = handlerWithModel.createMessage(systemPrompt, messages)
+		await messageGenerator.next()
+
+		expect(mockCreate).toHaveBeenCalledWith(
+			expect.objectContaining({
+				model: abliterationDefaultModelId,
+				max_tokens: modelInfo.maxTokens,
+				temperature: 0,
+				messages: expect.arrayContaining([{ role: "system", content: systemPrompt }]),
+				stream: true,
+				stream_options: { include_usage: true },
+			}),
+			undefined,
+		)
+	})
+})

src/api/providers/abliteration.ts

@@ -0,0 +1,18 @@
+import { type AbliterationModelId, abliterationDefaultModelId, abliterationModels } from "@roo-code/types"
+
+import type { ApiHandlerOptions } from "../../shared/api"
+
+import { BaseOpenAiCompatibleProvider } from "./base-openai-compatible-provider"
+
+export class AbliterationHandler extends BaseOpenAiCompatibleProvider<AbliterationModelId> {
+	constructor(options: ApiHandlerOptions) {
+		super({
+			...options,
+			providerName: "abliteration.ai",
+			baseURL: "https://api.abliteration.ai/v1",
+			apiKey: options.abliterationApiKey,
+			defaultProviderModelId: abliterationDefaultModelId,
+			providerModels: abliterationModels,
+		})
+	}
+}

src/api/providers/index.ts

@@ -1,3 +1,4 @@
+export { AbliterationHandler } from "./abliteration"
 export { AnthropicVertexHandler } from "./anthropic-vertex"
 export { AnthropicHandler } from "./anthropic"
 export { AwsBedrockHandler } from "./bedrock"

src/shared/ProfileValidator.ts

@@ -53,6 +53,7 @@ export class ProfileValidator {
 		switch (profile.apiProvider) {
 			case "openai":
 				return profile.openAiModelId
+			case "abliteration":
 			case "anthropic":
 			case "openai-native":
 			case "bedrock":

src/shared/__tests__/ProfileValidator.spec.ts

@@ -168,6 +168,7 @@ describe("ProfileValidator", () => {
 
 		// Test specific providers that use apiModelId
 		const apiModelProviders = [
+			"abliteration",
 			"anthropic",
 			"openai-native",
 			"bedrock",

webview-ui/src/components/settings/ApiOptions.tsx

@@ -9,6 +9,7 @@ import {
 	type ProviderSettings,
 	isRetiredProvider,
 	DEFAULT_CONSECUTIVE_MISTAKE_LIMIT,
+	abliterationDefaultModelId,
 	openRouterDefaultModelId,
 	poeDefaultModelId,
 	requestyDefaultModelId,
@@ -94,6 +95,7 @@ import {
 	Fireworks,
 	VercelAiGateway,
 	MiniMax,
+	Abliteration,
 } from "./providers"
 
 import { MODELS_BY_PROVIDER, PROVIDERS } from "./constants"
@@ -362,6 +364,7 @@ const ApiOptions = ({
 				fireworks: { field: "apiModelId", default: fireworksDefaultModelId },
 				poe: { field: "apiModelId", default: poeDefaultModelId },
 				roo: { field: "apiModelId", default: rooDefaultModelId },
+				abliteration: { field: "apiModelId", default: abliterationDefaultModelId },
 				"vercel-ai-gateway": { field: "vercelAiGatewayModelId", default: vercelAiGatewayDefaultModelId },
 				openai: { field: "openAiModelId" },
 				ollama: { field: "ollamaModelId" },
@@ -546,6 +549,13 @@ const ApiOptions = ({
 						/>
 					)}
 
+					{selectedProvider === "abliteration" && (
+						<Abliteration
+							apiConfiguration={apiConfiguration}
+							setApiConfigurationField={setApiConfigurationField}
+						/>
+					)}
+
 					{selectedProvider === "openai-codex" && (
 						<OpenAICodex
 							apiConfiguration={apiConfiguration}