@voquill/voice-ai 0.1.0

package/dist/claude.utils.js

@@ -0,0 +1,229 @@
+import Anthropic from "@anthropic-ai/sdk";
+import { retry, countWords } from "@voquill/utilities";
+export const CLAUDE_MODELS = [
+    "claude-opus-4-5-20251101",
+    "claude-opus-4-5",
+    "claude-3-7-sonnet-latest",
+    "claude-3-7-sonnet-20250219",
+    "claude-3-5-haiku-latest",
+    "claude-3-5-haiku-20241022",
+    "claude-haiku-4-5",
+    "claude-haiku-4-5-20251001",
+    "claude-sonnet-4-20250514",
+    "claude-sonnet-4-0",
+    "claude-4-sonnet-20250514",
+    "claude-sonnet-4-5",
+    "claude-sonnet-4-5-20250929",
+    "claude-opus-4-0",
+    "claude-opus-4-20250514",
+    "claude-4-opus-20250514",
+    "claude-opus-4-1-20250805",
+    "claude-3-opus-latest",
+    "claude-3-opus-20240229",
+    "claude-3-haiku-20240307",
+];
+const createClient = (apiKey) => {
+    return new Anthropic({
+        apiKey: apiKey.trim(),
+        dangerouslyAllowBrowser: true,
+    });
+};
+export const claudeGenerateTextResponse = async ({ apiKey, model = "claude-sonnet-4-20250514", system, prompt, jsonResponse, }) => {
+    return retry({
+        retries: 3,
+        fn: async () => {
+            const client = createClient(apiKey);
+            let finalPrompt = prompt;
+            if (jsonResponse) {
+                finalPrompt = `${prompt}\n\nRespond with valid JSON matching this schema: ${JSON.stringify(jsonResponse.schema)}`;
+            }
+            const response = await client.messages.create({
+                model,
+                max_tokens: 1024,
+                system: system ?? undefined,
+                messages: [{ role: "user", content: finalPrompt }],
+            });
+            console.log("claude llm usage:", response.usage);
+            const textBlock = response.content.find((block) => block.type === "text");
+            if (!textBlock || textBlock.type !== "text") {
+                throw new Error("No text response from Claude");
+            }
+            const content = textBlock.text;
+            const tokensUsed = (response.usage?.input_tokens ?? 0) +
+                (response.usage?.output_tokens ?? 0);
+            return {
+                text: content,
+                tokensUsed: tokensUsed || countWords(content),
+            };
+        },
+    });
+};
+export const claudeTestIntegration = async ({ apiKey, }) => {
+    const client = createClient(apiKey);
+    const response = await client.messages.create({
+        model: "claude-3-haiku-20240307",
+        max_tokens: 32,
+        messages: [
+            {
+                role: "user",
+                content: 'Reply with the single word "Hello."',
+            },
+        ],
+    });
+    const textBlock = response.content.find((block) => block.type === "text");
+    if (!textBlock || textBlock.type !== "text") {
+        throw new Error("No text response from Claude");
+    }
+    return textBlock.text.toLowerCase().includes("hello");
+};
+// ============================================================================
+// Streaming Chat
+// ============================================================================
+function llmMessagesToClaude(messages) {
+    let system;
+    const out = [];
+    for (const msg of messages) {
+        if (msg.role === "system") {
+            system = msg.content;
+            continue;
+        }
+        if (msg.role === "user") {
+            out.push({ role: "user", content: msg.content });
+            continue;
+        }
+        if (msg.role === "assistant") {
+            const content = [];
+            if (msg.content) {
+                content.push({ type: "text", text: msg.content });
+            }
+            for (const tc of msg.toolCalls ?? []) {
+                let parsedInput;
+                try {
+                    parsedInput = JSON.parse(tc.arguments);
+                }
+                catch {
+                    parsedInput = {};
+                }
+                content.push({
+                    type: "tool_use",
+                    id: tc.id,
+                    name: tc.name,
+                    input: parsedInput,
+                });
+            }
+            if (content.length > 0) {
+                out.push({ role: "assistant", content });
+            }
+            continue;
+        }
+        if (msg.role === "tool") {
+            out.push({
+                role: "user",
+                content: [
+                    {
+                        type: "tool_result",
+                        tool_use_id: msg.toolCallId,
+                        content: msg.content,
+                    },
+                ],
+            });
+        }
+    }
+    return { system, messages: out };
+}
+function claudeFinishReason(raw) {
+    switch (raw) {
+        case "end_turn":
+            return "stop";
+        case "max_tokens":
+            return "length";
+        case "tool_use":
+            return "tool-calls";
+        default:
+            return "other";
+    }
+}
+export async function* claudeStreamChat({ apiKey, model, input, }) {
+    const client = createClient(apiKey);
+    const { system, messages } = llmMessagesToClaude(input.messages);
+    const tools = input.tools && input.tools.length > 0
+        ? input.tools.map((t) => ({
+            name: t.name,
+            description: t.description ?? "",
+            input_schema: (t.parameters ?? {
+                type: "object",
+                properties: {},
+            }),
+        }))
+        : undefined;
+    let toolChoice;
+    if (input.toolChoice && tools) {
+        if (typeof input.toolChoice === "string") {
+            switch (input.toolChoice) {
+                case "auto":
+                    toolChoice = { type: "auto" };
+                    break;
+                case "required":
+                    toolChoice = { type: "any" };
+                    break;
+                case "none":
+                    toolChoice = undefined;
+                    break;
+            }
+        }
+        else {
+            toolChoice = { type: "tool", name: input.toolChoice.name };
+        }
+    }
+    const stream = client.messages.stream({
+        model,
+        max_tokens: input.maxTokens ?? 4096,
+        system,
+        messages,
+        tools,
+        tool_choice: toolChoice,
+        temperature: input.temperature,
+        top_p: input.topP,
+        stop_sequences: input.stopSequences,
+    });
+    const pendingToolCalls = [];
+    for await (const event of stream) {
+        if (event.type === "content_block_delta" &&
+            event.delta.type === "text_delta") {
+            yield { type: "text-delta", text: event.delta.text };
+        }
+        if (event.type === "content_block_delta" &&
+            event.delta.type === "input_json_delta") {
+            const last = pendingToolCalls[pendingToolCalls.length - 1];
+            if (last) {
+                last.arguments += event.delta.partial_json;
+            }
+        }
+        if (event.type === "content_block_start" &&
+            event.content_block.type === "tool_use") {
+            pendingToolCalls.push({
+                id: event.content_block.id,
+                name: event.content_block.name,
+                arguments: "",
+            });
+        }
+    }
+    for (const tc of pendingToolCalls) {
+        yield {
+            type: "tool-call",
+            id: tc.id,
+            name: tc.name,
+            arguments: tc.arguments,
+        };
+    }
+    const finalMessage = await stream.finalMessage();
+    yield {
+        type: "finish",
+        finishReason: claudeFinishReason(finalMessage.stop_reason),
+        usage: {
+            promptTokens: finalMessage.usage?.input_tokens,
+            completionTokens: finalMessage.usage?.output_tokens,
+        },
+        modelId: finalMessage.model,
+    };
+}

package/dist/deepgram.utils.d.ts

@@ -0,0 +1,5 @@
+export type DeepgramTestIntegrationArgs = {
+    apiKey: string;
+};
+export declare const deepgramTestIntegration: ({ apiKey, }: DeepgramTestIntegrationArgs) => Promise<boolean>;
+//# sourceMappingURL=deepgram.utils.d.ts.map

package/dist/deepgram.utils.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"deepgram.utils.d.ts","sourceRoot":"","sources":["../src/deepgram.utils.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,2BAA2B,GAAG;IACxC,MAAM,EAAE,MAAM,CAAC;CAChB,CAAC;AAEF,eAAO,MAAM,uBAAuB,GAAI,aAErC,2BAA2B,KAAG,OAAO,CAAC,OAAO,CA4B/C,CAAC"}

package/dist/deepgram.utils.js

@@ -0,0 +1,25 @@
+export const deepgramTestIntegration = ({ apiKey, }) => {
+    return new Promise((resolve) => {
+        const wsUrl = "wss://api.deepgram.com/v1/listen?encoding=linear16&sample_rate=16000&model=nova-3";
+        const ws = new WebSocket(wsUrl, ["token", apiKey]);
+        const timeout = setTimeout(() => {
+            ws.close();
+            resolve(false);
+        }, 5000);
+        ws.onopen = () => {
+            clearTimeout(timeout);
+            ws.close();
+            resolve(true);
+        };
+        ws.onerror = () => {
+            clearTimeout(timeout);
+            resolve(false);
+        };
+        ws.onclose = (event) => {
+            clearTimeout(timeout);
+            if (event.code === 1008 || event.code === 4001 || event.code === 4003) {
+                resolve(false);
+            }
+        };
+    });
+};

package/dist/deepseek.utils.d.ts

@@ -0,0 +1,26 @@
+import type { JsonResponse, LlmChatInput, LlmStreamEvent } from "@voquill/types";
+export declare const DEEPSEEK_MODELS: readonly ["deepseek-chat", "deepseek-reasoner"];
+export type DeepseekModel = (typeof DEEPSEEK_MODELS)[number];
+export type DeepseekGenerateTextArgs = {
+    apiKey: string;
+    model?: DeepseekModel;
+    system?: string;
+    prompt: string;
+    jsonResponse?: JsonResponse;
+};
+export type DeepseekGenerateResponseOutput = {
+    text: string;
+    tokensUsed: number;
+};
+export declare const deepseekGenerateTextResponse: ({ apiKey, model, system, prompt, jsonResponse, }: DeepseekGenerateTextArgs) => Promise<DeepseekGenerateResponseOutput>;
+export type DeepseekTestIntegrationArgs = {
+    apiKey: string;
+};
+export declare const deepseekTestIntegration: ({ apiKey, }: DeepseekTestIntegrationArgs) => Promise<boolean>;
+export type DeepseekStreamChatArgs = {
+    apiKey: string;
+    model: string;
+    input: LlmChatInput;
+};
+export declare function deepseekStreamChat({ apiKey, model, input, }: DeepseekStreamChatArgs): AsyncGenerator<LlmStreamEvent>;
+//# sourceMappingURL=deepseek.utils.d.ts.map

package/dist/deepseek.utils.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"deepseek.utils.d.ts","sourceRoot":"","sources":["../src/deepseek.utils.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,YAAY,EAAE,YAAY,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAGjF,eAAO,MAAM,eAAe,iDAAkD,CAAC;AAC/E,MAAM,MAAM,aAAa,GAAG,CAAC,OAAO,eAAe,CAAC,CAAC,MAAM,CAAC,CAAC;AAkC7D,MAAM,MAAM,wBAAwB,GAAG;IACrC,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,aAAa,CAAC;IACtB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,CAAC,EAAE,YAAY,CAAC;CAC7B,CAAC;AAEF,MAAM,MAAM,8BAA8B,GAAG;IAC3C,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,MAAM,CAAC;CACpB,CAAC;AAEF,eAAO,MAAM,4BAA4B,GAAU,kDAMhD,wBAAwB,KAAG,OAAO,CAAC,8BAA8B,CA8CnE,CAAC;AAEF,MAAM,MAAM,2BAA2B,GAAG;IACxC,MAAM,EAAE,MAAM,CAAC;CAChB,CAAC;AAEF,eAAO,MAAM,uBAAuB,GAAU,aAE3C,2BAA2B,KAAG,OAAO,CAAC,OAAO,CAgC/C,CAAC;AAMF,MAAM,MAAM,sBAAsB,GAAG;IACnC,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,YAAY,CAAC;CACrB,CAAC;AAEF,wBAAuB,kBAAkB,CAAC,EACxC,MAAM,EACN,KAAK,EACL,KAAK,GACN,EAAE,sBAAsB,GAAG,cAAc,CAAC,cAAc,CAAC,CAGzD"}

package/dist/deepseek.utils.js

@@ -0,0 +1,102 @@
+import OpenAI from "openai";
+import { retry, countWords } from "@voquill/utilities";
+import { openaiCompatibleStreamChat } from "./openai.utils";
+export const DEEPSEEK_MODELS = ["deepseek-chat", "deepseek-reasoner"];
+const DEEPSEEK_BASE_URL = "https://api.deepseek.com";
+const contentToString = (content) => {
+    if (!content) {
+        return "";
+    }
+    if (typeof content === "string") {
+        return content;
+    }
+    return content
+        .map((part) => {
+        if (part.type === "text") {
+            return part.text ?? "";
+        }
+        return "";
+    })
+        .join("")
+        .trim();
+};
+const createClient = (apiKey) => {
+    return new OpenAI({
+        apiKey: apiKey.trim(),
+        baseURL: DEEPSEEK_BASE_URL,
+        dangerouslyAllowBrowser: true, // This is safe because Voquill natively on desktop
+    });
+};
+export const deepseekGenerateTextResponse = async ({ apiKey, model = "deepseek-chat", system, prompt, jsonResponse, }) => {
+    return retry({
+        retries: 3,
+        fn: async () => {
+            const client = createClient(apiKey);
+            const messages = [];
+            if (system) {
+                messages.push({ role: "system", content: system });
+            }
+            let finalPrompt = prompt;
+            if (jsonResponse) {
+                finalPrompt = `${prompt}\n\nRespond with valid JSON matching this schema: ${JSON.stringify(jsonResponse.schema)}`;
+            }
+            const userParts = [];
+            userParts.push({ type: "text", text: finalPrompt });
+            messages.push({ role: "user", content: userParts });
+            const response = await client.chat.completions.create({
+                messages,
+                model,
+                temperature: 1,
+                max_tokens: 1024,
+                top_p: 1,
+                response_format: jsonResponse ? { type: "json_object" } : undefined,
+            });
+            console.log("deepseek llm usage:", response.usage);
+            if (!response.choices || response.choices.length === 0) {
+                throw new Error("No response from DeepSeek");
+            }
+            const result = response.choices[0].message.content;
+            if (!result) {
+                throw new Error("Content is empty");
+            }
+            const content = contentToString(result);
+            return {
+                text: content,
+                tokensUsed: response.usage?.total_tokens ?? countWords(content),
+            };
+        },
+    });
+};
+export const deepseekTestIntegration = async ({ apiKey, }) => {
+    const client = createClient(apiKey);
+    const response = await client.chat.completions.create({
+        messages: [
+            {
+                role: "user",
+                content: [
+                    {
+                        type: "text",
+                        text: `Reply with the single word "Hello."`,
+                    },
+                ],
+            },
+        ],
+        model: "deepseek-chat",
+        temperature: 0,
+        max_tokens: 32,
+        top_p: 1,
+    });
+    if (!response.choices || response.choices.length === 0) {
+        throw new Error("No response from DeepSeek");
+    }
+    const first = response.choices[0];
+    const content = contentToString(first?.message?.content);
+    if (!content) {
+        throw new Error("Response content is empty");
+    }
+    return content.toLowerCase().includes("hello");
+};
+export async function* deepseekStreamChat({ apiKey, model, input, }) {
+    const client = createClient(apiKey);
+    yield* openaiCompatibleStreamChat(client, model, input);
+}

package/dist/elevenlabs.utils.d.ts

@@ -0,0 +1,6 @@
+export type ElevenLabsTestIntegrationArgs = {
+    apiKey: string;
+};
+export declare const elevenlabsTestIntegration: ({ apiKey, }: ElevenLabsTestIntegrationArgs) => Promise<boolean>;
+export declare const convertFloat32ToBase64PCM16: (float32Array: Float32Array | number[]) => string;
+//# sourceMappingURL=elevenlabs.utils.d.ts.map

package/dist/elevenlabs.utils.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"elevenlabs.utils.d.ts","sourceRoot":"","sources":["../src/elevenlabs.utils.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,6BAA6B,GAAG;IAC1C,MAAM,EAAE,MAAM,CAAC;CAChB,CAAC;AAEF,eAAO,MAAM,yBAAyB,GAAU,aAE7C,6BAA6B,KAAG,OAAO,CAAC,OAAO,CAUjD,CAAC;AAEF,eAAO,MAAM,2BAA2B,GACtC,cAAc,YAAY,GAAG,MAAM,EAAE,KACpC,MAmBF,CAAC"}

package/dist/elevenlabs.utils.js

@@ -0,0 +1,29 @@
+export const elevenlabsTestIntegration = async ({ apiKey, }) => {
+    try {
+        const response = await fetch("https://api.elevenlabs.io/v1/user", {
+            method: "GET",
+            headers: { "xi-api-key": apiKey },
+        });
+        return response.ok;
+    }
+    catch {
+        return false;
+    }
+};
+export const convertFloat32ToBase64PCM16 = (float32Array) => {
+    const samples = Array.isArray(float32Array)
+        ? float32Array
+        : Array.from(float32Array);
+    const buffer = new ArrayBuffer(samples.length * 2);
+    const view = new DataView(buffer);
+    for (let i = 0; i < samples.length; i++) {
+        const s = Math.max(-1, Math.min(1, samples[i]));
+        view.setInt16(i * 2, s < 0 ? s * 0x8000 : s * 0x7fff, true);
+    }
+    const bytes = new Uint8Array(buffer);
+    let binary = "";
+    for (let i = 0; i < bytes.length; i++) {
+        binary += String.fromCharCode(bytes[i]);
+    }
+    return btoa(binary);
+};

package/dist/gemini.utils.d.ts

@@ -0,0 +1,41 @@
+import type { JsonResponse, LlmChatInput, LlmStreamEvent } from "@voquill/types";
+export declare const GEMINI_GENERATE_TEXT_MODELS: readonly ["gemini-2.5-flash", "gemini-2.5-pro", "gemini-3-flash-preview", "gemini-3-pro-preview", "gemini-2.5-flash-lite"];
+export type GeminiGenerateTextModel = (typeof GEMINI_GENERATE_TEXT_MODELS)[number];
+export declare const GEMINI_TRANSCRIPTION_MODELS: readonly ["gemini-2.5-flash", "gemini-2.5-pro", "gemini-3-flash-preview"];
+export type GeminiTranscriptionModel = (typeof GEMINI_TRANSCRIPTION_MODELS)[number];
+export type GeminiTranscriptionArgs = {
+    apiKey: string;
+    model?: GeminiTranscriptionModel;
+    blob: ArrayBuffer | Buffer;
+    mimeType?: string;
+    prompt?: string;
+    language?: string;
+};
+export type GeminiTranscribeAudioOutput = {
+    text: string;
+    wordsUsed: number;
+};
+export declare const geminiTranscribeAudio: ({ apiKey, model, blob, mimeType, prompt, language, }: GeminiTranscriptionArgs) => Promise<GeminiTranscribeAudioOutput>;
+export type GeminiGenerateTextArgs = {
+    apiKey: string;
+    model?: GeminiGenerateTextModel;
+    system?: string;
+    prompt: string;
+    jsonResponse?: JsonResponse;
+};
+export type GeminiGenerateResponseOutput = {
+    text: string;
+    tokensUsed: number;
+};
+export declare const geminiGenerateTextResponse: ({ apiKey, model, system, prompt, jsonResponse, }: GeminiGenerateTextArgs) => Promise<GeminiGenerateResponseOutput>;
+export type GeminiTestIntegrationArgs = {
+    apiKey: string;
+};
+export declare const geminiTestIntegration: ({ apiKey, }: GeminiTestIntegrationArgs) => Promise<boolean>;
+export type GeminiStreamChatArgs = {
+    apiKey: string;
+    model: string;
+    input: LlmChatInput;
+};
+export declare function geminiStreamChat({ apiKey, model, input, }: GeminiStreamChatArgs): AsyncGenerator<LlmStreamEvent>;
+//# sourceMappingURL=gemini.utils.d.ts.map

package/dist/gemini.utils.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"gemini.utils.d.ts","sourceRoot":"","sources":["../src/gemini.utils.ts"],"names":[],"mappings":"AAQA,OAAO,KAAK,EACV,YAAY,EACZ,YAAY,EAGZ,cAAc,EACf,MAAM,gBAAgB,CAAC;AAExB,eAAO,MAAM,2BAA2B,4HAM9B,CAAC;AACX,MAAM,MAAM,uBAAuB,GACjC,CAAC,OAAO,2BAA2B,CAAC,CAAC,MAAM,CAAC,CAAC;AAE/C,eAAO,MAAM,2BAA2B,2EAI9B,CAAC;AACX,MAAM,MAAM,wBAAwB,GAClC,CAAC,OAAO,2BAA2B,CAAC,CAAC,MAAM,CAAC,CAAC;AAgD/C,MAAM,MAAM,uBAAuB,GAAG;IACpC,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,wBAAwB,CAAC;IACjC,IAAI,EAAE,WAAW,GAAG,MAAM,CAAC;IAC3B,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB,CAAC;AAEF,MAAM,MAAM,2BAA2B,GAAG;IACxC,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;CACnB,CAAC;AAEF,eAAO,MAAM,qBAAqB,GAAU,sDAOzC,uBAAuB,KAAG,OAAO,CAAC,2BAA2B,CA0C/D,CAAC;AAEF,MAAM,MAAM,sBAAsB,GAAG;IACnC,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,uBAAuB,CAAC;IAChC,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,CAAC,EAAE,YAAY,CAAC;CAC7B,CAAC;AAEF,MAAM,MAAM,4BAA4B,GAAG;IACzC,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,MAAM,CAAC;CACpB,CAAC;AAEF,eAAO,MAAM,0BAA0B,GAAU,kDAM9C,sBAAsB,KAAG,OAAO,CAAC,4BAA4B,CA4C/D,CAAC;AAEF,MAAM,MAAM,yBAAyB,GAAG;IACtC,MAAM,EAAE,MAAM,CAAC;CAChB,CAAC;AAEF,eAAO,MAAM,qBAAqB,GAAU,aAEzC,yBAAyB,KAAG,OAAO,CAAC,OAAO,CAc7C,CAAC;AA6EF,MAAM,MAAM,oBAAoB,GAAG;IACjC,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,YAAY,CAAC;CACrB,CAAC;AAEF,wBAAuB,gBAAgB,CAAC,EACtC,MAAM,EACN,KAAK,EACL,KAAK,GACN,EAAE,oBAAoB,GAAG,cAAc,CAAC,cAAc,CAAC,CAuFvD"}