@oh-my-pi/pi-ai 3.15.0 → 3.20.0

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@oh-my-pi/pi-ai",
-	"version": "3.15.0",
+	"version": "3.20.0",
 	"description": "Unified LLM API with automatic model discovery and provider configuration",
 	"type": "module",
 	"main": "./src/index.ts",

package/src/bun-imports.d.ts

@@ -0,0 +1,14 @@
+/**
+ * Type declarations for Bun's import attributes.
+ * These allow importing non-JS files as text at build time.
+ */
+
+declare module "*.md" {
+	const content: string;
+	export default content;
+}
+
+declare module "*.txt" {
+	const content: string;
+	export default content;
+}

package/src/cli.ts

@@ -6,6 +6,7 @@ import { loginGitHubCopilot } from "./utils/oauth/github-copilot";
 import { loginAntigravity } from "./utils/oauth/google-antigravity";
 import { loginGeminiCli } from "./utils/oauth/google-gemini-cli";
 import { getOAuthProviders } from "./utils/oauth/index";
+import { loginOpenAICodex } from "./utils/oauth/openai-codex";
 import type { OAuthCredentials, OAuthProvider } from "./utils/oauth/types";
 
 const AUTH_FILE = "auth.json";
@@ -89,6 +90,19 @@ async function login(provider: OAuthProvider): Promise<void> {
 				(msg) => console.log(msg),
 			);
 			break;
+		case "openai-codex":
+			credentials = await loginOpenAICodex({
+				onAuth: (info) => {
+					console.log(`\nOpen this URL in your browser:\n${info.url}`);
+					if (info.instructions) console.log(info.instructions);
+					console.log();
+				},
+				onPrompt: async (p) => {
+					return await promptFn(`${p.message}${p.placeholder ? ` (${p.placeholder})` : ""}:`);
+				},
+				onProgress: (msg) => console.log(msg),
+			});
+			break;
 	}
 
 	const auth = await loadAuth();
@@ -114,6 +128,7 @@ Providers:
   github-copilot    GitHub Copilot
   google-gemini-cli Google Gemini CLI
   google-antigravity Antigravity (Gemini 3, Claude, GPT-OSS)
+  openai-codex      OpenAI Codex (ChatGPT Plus/Pro)
 
 Examples:
   npx @oh-my-pi/pi-ai login              # interactive provider selection
@@ -141,7 +156,7 @@ Examples:
 			}
 			console.log();
 
-			const choice = await prompt("Enter number (1-4): ");
+			const choice = await prompt(`Enter number (1-${PROVIDERS.length}): `);
 
 			const index = parseInt(choice, 10) - 1;
 			if (index < 0 || index >= PROVIDERS.length) {

package/src/index.ts

@@ -1,3 +1,5 @@
+/// <reference path="./bun-imports.d.ts" />
+
 export * from "./models";
 export * from "./providers/anthropic";
 export * from "./providers/google";

package/src/models.generated.ts

@@ -2755,23 +2755,6 @@ export const MODELS = {
 			contextWindow: 200000,
 			maxTokens: 4096,
 		} satisfies Model<"openai-completions">,
-		"anthropic/claude-3-opus": {
-			id: "anthropic/claude-3-opus",
-			name: "Anthropic: Claude 3 Opus",
-			api: "openai-completions",
-			provider: "openrouter",
-			baseUrl: "https://openrouter.ai/api/v1",
-			reasoning: false,
-			input: ["text", "image"],
-			cost: {
-				input: 15,
-				output: 75,
-				cacheRead: 1.5,
-				cacheWrite: 18.75,
-			},
-			contextWindow: 200000,
-			maxTokens: 4096,
-		} satisfies Model<"openai-completions">,
 		"anthropic/claude-3.5-haiku": {
 			id: "anthropic/claude-3.5-haiku",
 			name: "Anthropic: Claude 3.5 Haiku",
@@ -3605,6 +3588,23 @@ export const MODELS = {
 			contextWindow: 128000,
 			maxTokens: 16384,
 		} satisfies Model<"openai-completions">,
+		"kwaipilot/kat-coder-pro": {
+			id: "kwaipilot/kat-coder-pro",
+			name: "Kwaipilot: KAT-Coder-Pro V1",
+			api: "openai-completions",
+			provider: "openrouter",
+			baseUrl: "https://openrouter.ai/api/v1",
+			reasoning: false,
+			input: ["text"],
+			cost: {
+				input: 0.207,
+				output: 0.828,
+				cacheRead: 0.0414,
+				cacheWrite: 0,
+			},
+			contextWindow: 256000,
+			maxTokens: 128000,
+		} satisfies Model<"openai-completions">,
 		"kwaipilot/kat-coder-pro:free": {
 			id: "kwaipilot/kat-coder-pro:free",
 			name: "Kwaipilot: KAT-Coder-Pro V1 (free)",
@@ -3620,7 +3620,7 @@ export const MODELS = {
 				cacheWrite: 0,
 			},
 			contextWindow: 256000,
-			maxTokens: 32768,
+			maxTokens: 128000,
 		} satisfies Model<"openai-completions">,
 		"meta-llama/llama-3-70b-instruct": {
 			id: "meta-llama/llama-3-70b-instruct",
@@ -3700,11 +3700,11 @@ export const MODELS = {
 			input: ["text"],
 			cost: {
 				input: 0.02,
-				output: 0.03,
+				output: 0.049999999999999996,
 				cacheRead: 0,
 				cacheWrite: 0,
 			},
-			contextWindow: 131072,
+			contextWindow: 16384,
 			maxTokens: 16384,
 		} satisfies Model<"openai-completions">,
 		"meta-llama/llama-3.2-3b-instruct": {
@@ -6240,6 +6240,23 @@ export const MODELS = {
 			contextWindow: 163840,
 			maxTokens: 65536,
 		} satisfies Model<"openai-completions">,
+		"tngtech/tng-r1t-chimera:free": {
+			id: "tngtech/tng-r1t-chimera:free",
+			name: "TNG: R1T Chimera (free)",
+			api: "openai-completions",
+			provider: "openrouter",
+			baseUrl: "https://openrouter.ai/api/v1",
+			reasoning: true,
+			input: ["text"],
+			cost: {
+				input: 0,
+				output: 0,
+				cacheRead: 0,
+				cacheWrite: 0,
+			},
+			contextWindow: 163840,
+			maxTokens: 65536,
+		} satisfies Model<"openai-completions">,
 		"x-ai/grok-3": {
 			id: "x-ai/grok-3",
 			name: "xAI: Grok 3",

package/src/models.ts

@@ -12,27 +12,34 @@ for (const [provider, models] of Object.entries(MODELS)) {
 	modelRegistry.set(provider, providerModels);
 }
 
+type ProviderModels = typeof MODELS;
+type ProviderWithModels = keyof ProviderModels;
+
 type ModelApi<
-	TProvider extends KnownProvider,
-	TModelId extends keyof (typeof MODELS)[TProvider],
-> = (typeof MODELS)[TProvider][TModelId] extends { api: infer TApi } ? (TApi extends Api ? TApi : never) : never;
+	TProvider extends ProviderWithModels,
+	TModelId extends keyof ProviderModels[TProvider],
+> = ProviderModels[TProvider][TModelId] extends { api: infer TApi } ? (TApi extends Api ? TApi : never) : never;
 
-export function getModel<TProvider extends KnownProvider, TModelId extends keyof (typeof MODELS)[TProvider]>(
+export function getModel<TProvider extends ProviderWithModels, TModelId extends keyof ProviderModels[TProvider]>(
 	provider: TProvider,
 	modelId: TModelId,
-): Model<ModelApi<TProvider, TModelId>> {
-	return modelRegistry.get(provider)?.get(modelId as string) as Model<ModelApi<TProvider, TModelId>>;
+): Model<ModelApi<TProvider, TModelId>>;
+export function getModel(provider: KnownProvider, modelId: string): Model<Api> | undefined;
+export function getModel(provider: KnownProvider, modelId: string): Model<Api> | undefined {
+	return modelRegistry.get(provider)?.get(modelId as string) as Model<Api> | undefined;
 }
 
 export function getProviders(): KnownProvider[] {
 	return Array.from(modelRegistry.keys()) as KnownProvider[];
 }
 
-export function getModels<TProvider extends KnownProvider>(
+export function getModels<TProvider extends ProviderWithModels>(
 	provider: TProvider,
-): Model<ModelApi<TProvider, keyof (typeof MODELS)[TProvider]>>[] {
+): Model<ModelApi<TProvider, keyof ProviderModels[TProvider]>>[];
+export function getModels(provider: KnownProvider): Model<Api>[];
+export function getModels(provider: KnownProvider): Model<Api>[] {
 	const models = modelRegistry.get(provider);
-	return models ? (Array.from(models.values()) as Model<ModelApi<TProvider, keyof (typeof MODELS)[TProvider]>>[]) : [];
+	return models ? (Array.from(models.values()) as Model<Api>[]) : [];
 }
 
 export function calculateCost<TApi extends Api>(model: Model<TApi>, usage: Usage): Usage["cost"] {

package/src/providers/google-shared.ts

@@ -7,7 +7,7 @@ import type { Context, ImageContent, Model, StopReason, TextContent, Tool } from
 import { sanitizeSurrogates } from "../utils/sanitize-unicode";
 import { transformMessages } from "./transorm-messages";
 
-type GoogleApiType = "google-generative-ai" | "google-gemini-cli";
+type GoogleApiType = "google-generative-ai" | "google-gemini-cli" | "google-vertex";
 
 /**
  * Convert internal messages to Gemini Content[] format.

package/src/providers/google-vertex.ts

@@ -0,0 +1,355 @@
+import {
+	type GenerateContentConfig,
+	type GenerateContentParameters,
+	GoogleGenAI,
+	type ThinkingConfig,
+	ThinkingLevel,
+} from "@google/genai";
+import { calculateCost } from "../models";
+import type {
+	Api,
+	AssistantMessage,
+	Context,
+	Model,
+	StreamFunction,
+	StreamOptions,
+	TextContent,
+	ThinkingContent,
+	ToolCall,
+} from "../types";
+import { AssistantMessageEventStream } from "../utils/event-stream";
+import { sanitizeSurrogates } from "../utils/sanitize-unicode";
+import type { GoogleThinkingLevel } from "./google-gemini-cli";
+import { convertMessages, convertTools, mapStopReason, mapToolChoice } from "./google-shared";
+
+export interface GoogleVertexOptions extends StreamOptions {
+	toolChoice?: "auto" | "none" | "any";
+	thinking?: {
+		enabled: boolean;
+		budgetTokens?: number; // -1 for dynamic, 0 to disable
+		level?: GoogleThinkingLevel;
+	};
+	project?: string;
+	location?: string;
+}
+
+const API_VERSION = "v1";
+
+const THINKING_LEVEL_MAP: Record<GoogleThinkingLevel, ThinkingLevel> = {
+	THINKING_LEVEL_UNSPECIFIED: ThinkingLevel.THINKING_LEVEL_UNSPECIFIED,
+	MINIMAL: ThinkingLevel.MINIMAL,
+	LOW: ThinkingLevel.LOW,
+	MEDIUM: ThinkingLevel.MEDIUM,
+	HIGH: ThinkingLevel.HIGH,
+};
+
+// Counter for generating unique tool call IDs
+let toolCallCounter = 0;
+
+export const streamGoogleVertex: StreamFunction<"google-vertex"> = (
+	model: Model<"google-vertex">,
+	context: Context,
+	options?: GoogleVertexOptions,
+): AssistantMessageEventStream => {
+	const stream = new AssistantMessageEventStream();
+
+	(async () => {
+		const output: AssistantMessage = {
+			role: "assistant",
+			content: [],
+			api: "google-vertex" as Api,
+			provider: model.provider,
+			model: model.id,
+			usage: {
+				input: 0,
+				output: 0,
+				cacheRead: 0,
+				cacheWrite: 0,
+				totalTokens: 0,
+				cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },
+			},
+			stopReason: "stop",
+			timestamp: Date.now(),
+		};
+
+		try {
+			const project = resolveProject(options);
+			const location = resolveLocation(options);
+			const client = createClient(model, project, location);
+			const params = buildParams(model, context, options);
+			const googleStream = await client.models.generateContentStream(params);
+
+			stream.push({ type: "start", partial: output });
+			let currentBlock: TextContent | ThinkingContent | null = null;
+			const blocks = output.content;
+			const blockIndex = () => blocks.length - 1;
+			for await (const chunk of googleStream) {
+				const candidate = chunk.candidates?.[0];
+				if (candidate?.content?.parts) {
+					for (const part of candidate.content.parts) {
+						if (part.text !== undefined) {
+							const isThinking = part.thought === true;
+							if (
+								!currentBlock ||
+								(isThinking && currentBlock.type !== "thinking") ||
+								(!isThinking && currentBlock.type !== "text")
+							) {
+								if (currentBlock) {
+									if (currentBlock.type === "text") {
+										stream.push({
+											type: "text_end",
+											contentIndex: blocks.length - 1,
+											content: currentBlock.text,
+											partial: output,
+										});
+									} else {
+										stream.push({
+											type: "thinking_end",
+											contentIndex: blockIndex(),
+											content: currentBlock.thinking,
+											partial: output,
+										});
+									}
+								}
+								if (isThinking) {
+									currentBlock = { type: "thinking", thinking: "", thinkingSignature: undefined };
+									output.content.push(currentBlock);
+									stream.push({ type: "thinking_start", contentIndex: blockIndex(), partial: output });
+								} else {
+									currentBlock = { type: "text", text: "" };
+									output.content.push(currentBlock);
+									stream.push({ type: "text_start", contentIndex: blockIndex(), partial: output });
+								}
+							}
+							if (currentBlock.type === "thinking") {
+								currentBlock.thinking += part.text;
+								currentBlock.thinkingSignature = part.thoughtSignature;
+								stream.push({
+									type: "thinking_delta",
+									contentIndex: blockIndex(),
+									delta: part.text,
+									partial: output,
+								});
+							} else {
+								currentBlock.text += part.text;
+								stream.push({
+									type: "text_delta",
+									contentIndex: blockIndex(),
+									delta: part.text,
+									partial: output,
+								});
+							}
+						}
+
+						if (part.functionCall) {
+							if (currentBlock) {
+								if (currentBlock.type === "text") {
+									stream.push({
+										type: "text_end",
+										contentIndex: blockIndex(),
+										content: currentBlock.text,
+										partial: output,
+									});
+								} else {
+									stream.push({
+										type: "thinking_end",
+										contentIndex: blockIndex(),
+										content: currentBlock.thinking,
+										partial: output,
+									});
+								}
+								currentBlock = null;
+							}
+
+							const providedId = part.functionCall.id;
+							const needsNewId =
+								!providedId || output.content.some((b) => b.type === "toolCall" && b.id === providedId);
+							const toolCallId = needsNewId
+								? `${part.functionCall.name}_${Date.now()}_${++toolCallCounter}`
+								: providedId;
+
+							const toolCall: ToolCall = {
+								type: "toolCall",
+								id: toolCallId,
+								name: part.functionCall.name || "",
+								arguments: part.functionCall.args as Record<string, any>,
+								...(part.thoughtSignature && { thoughtSignature: part.thoughtSignature }),
+							};
+
+							output.content.push(toolCall);
+							stream.push({ type: "toolcall_start", contentIndex: blockIndex(), partial: output });
+							stream.push({
+								type: "toolcall_delta",
+								contentIndex: blockIndex(),
+								delta: JSON.stringify(toolCall.arguments),
+								partial: output,
+							});
+							stream.push({ type: "toolcall_end", contentIndex: blockIndex(), toolCall, partial: output });
+						}
+					}
+				}
+
+				if (candidate?.finishReason) {
+					output.stopReason = mapStopReason(candidate.finishReason);
+					if (output.content.some((b) => b.type === "toolCall")) {
+						output.stopReason = "toolUse";
+					}
+				}
+
+				if (chunk.usageMetadata) {
+					output.usage = {
+						input: chunk.usageMetadata.promptTokenCount || 0,
+						output:
+							(chunk.usageMetadata.candidatesTokenCount || 0) + (chunk.usageMetadata.thoughtsTokenCount || 0),
+						cacheRead: chunk.usageMetadata.cachedContentTokenCount || 0,
+						cacheWrite: 0,
+						totalTokens: chunk.usageMetadata.totalTokenCount || 0,
+						cost: {
+							input: 0,
+							output: 0,
+							cacheRead: 0,
+							cacheWrite: 0,
+							total: 0,
+						},
+					};
+					calculateCost(model, output.usage);
+				}
+			}
+
+			if (currentBlock) {
+				if (currentBlock.type === "text") {
+					stream.push({
+						type: "text_end",
+						contentIndex: blockIndex(),
+						content: currentBlock.text,
+						partial: output,
+					});
+				} else {
+					stream.push({
+						type: "thinking_end",
+						contentIndex: blockIndex(),
+						content: currentBlock.thinking,
+						partial: output,
+					});
+				}
+			}
+
+			if (options?.signal?.aborted) {
+				throw new Error("Request was aborted");
+			}
+
+			if (output.stopReason === "aborted" || output.stopReason === "error") {
+				throw new Error("An unknown error occurred");
+			}
+
+			stream.push({ type: "done", reason: output.stopReason, message: output });
+			stream.end();
+		} catch (error) {
+			// Remove internal index property used during streaming
+			for (const block of output.content) {
+				if ("index" in block) {
+					delete (block as { index?: number }).index;
+				}
+			}
+			output.stopReason = options?.signal?.aborted ? "aborted" : "error";
+			output.errorMessage = error instanceof Error ? error.message : JSON.stringify(error);
+			stream.push({ type: "error", reason: output.stopReason, error: output });
+			stream.end();
+		}
+	})();
+
+	return stream;
+};
+
+function createClient(model: Model<"google-vertex">, project: string, location: string): GoogleGenAI {
+	const httpOptions: { headers?: Record<string, string> } = {};
+
+	if (model.headers) {
+		httpOptions.headers = { ...model.headers };
+	}
+
+	const hasHttpOptions = Object.values(httpOptions).some(Boolean);
+
+	return new GoogleGenAI({
+		vertexai: true,
+		project,
+		location,
+		apiVersion: API_VERSION,
+		httpOptions: hasHttpOptions ? httpOptions : undefined,
+	});
+}
+
+function resolveProject(options?: GoogleVertexOptions): string {
+	const project = options?.project || process.env.GOOGLE_CLOUD_PROJECT || process.env.GCLOUD_PROJECT;
+	if (!project) {
+		throw new Error(
+			"Vertex AI requires a project ID. Set GOOGLE_CLOUD_PROJECT/GCLOUD_PROJECT or pass project in options.",
+		);
+	}
+	return project;
+}
+
+function resolveLocation(options?: GoogleVertexOptions): string {
+	const location = options?.location || process.env.GOOGLE_CLOUD_LOCATION;
+	if (!location) {
+		throw new Error("Vertex AI requires a location. Set GOOGLE_CLOUD_LOCATION or pass location in options.");
+	}
+	return location;
+}
+
+function buildParams(
+	model: Model<"google-vertex">,
+	context: Context,
+	options: GoogleVertexOptions = {},
+): GenerateContentParameters {
+	const contents = convertMessages(model, context);
+
+	const generationConfig: GenerateContentConfig = {};
+	if (options.temperature !== undefined) {
+		generationConfig.temperature = options.temperature;
+	}
+	if (options.maxTokens !== undefined) {
+		generationConfig.maxOutputTokens = options.maxTokens;
+	}
+
+	const config: GenerateContentConfig = {
+		...(Object.keys(generationConfig).length > 0 && generationConfig),
+		...(context.systemPrompt && { systemInstruction: sanitizeSurrogates(context.systemPrompt) }),
+		...(context.tools && context.tools.length > 0 && { tools: convertTools(context.tools) }),
+	};
+
+	if (context.tools && context.tools.length > 0 && options.toolChoice) {
+		config.toolConfig = {
+			functionCallingConfig: {
+				mode: mapToolChoice(options.toolChoice),
+			},
+		};
+	} else {
+		config.toolConfig = undefined;
+	}
+
+	if (options.thinking?.enabled && model.reasoning) {
+		const thinkingConfig: ThinkingConfig = { includeThoughts: true };
+		if (options.thinking.level !== undefined) {
+			thinkingConfig.thinkingLevel = THINKING_LEVEL_MAP[options.thinking.level];
+		} else if (options.thinking.budgetTokens !== undefined) {
+			thinkingConfig.thinkingBudget = options.thinking.budgetTokens;
+		}
+		config.thinkingConfig = thinkingConfig;
+	}
+
+	if (options.signal) {
+		if (options.signal.aborted) {
+			throw new Error("Request aborted");
+		}
+		config.abortSignal = options.signal;
+	}
+
+	const params: GenerateContentParameters = {
+		model: model.id,
+		contents,
+		config,
+	};
+
+	return params;
+}

package/src/providers/openai-codex/constants.ts

@@ -0,0 +1,25 @@
+/**
+ * Constants for OpenAI Codex (ChatGPT OAuth) backend
+ */
+
+export const CODEX_BASE_URL = "https://chatgpt.com/backend-api";
+
+export const OPENAI_HEADERS = {
+	BETA: "OpenAI-Beta",
+	ACCOUNT_ID: "chatgpt-account-id",
+	ORIGINATOR: "originator",
+	SESSION_ID: "session_id",
+	CONVERSATION_ID: "conversation_id",
+} as const;
+
+export const OPENAI_HEADER_VALUES = {
+	BETA_RESPONSES: "responses=experimental",
+	ORIGINATOR_CODEX: "codex_cli_rs",
+} as const;
+
+export const URL_PATHS = {
+	RESPONSES: "/responses",
+	CODEX_RESPONSES: "/codex/responses",
+} as const;
+
+export const JWT_CLAIM_PATH = "https://api.openai.com/auth" as const;