App & tools

WebVoice is both: a ready-to-use web application (voice memos, TTS, STT, chat, images, billing) and a set of HTTP endpoints you can embed anywhere you can run HTTPS requests — mobile apps, backends, intranets, e-commerce, or internal dashboards.

API forward — how it works

Point your HTTP client or OpenAI SDK at WebVoice instead of OpenAI: we validate your API key, enforce model access and credits, forward the request to the configured provider, and return an OpenAI-shaped JSON (or SSE stream). You do not need Groq/DeepSeek/OpenRouter keys in your app.

• Base URL: https://webvoice.easytaskflow.app/api/v1/ — full API reference
• Auth: Header X-API-Key or Authorization: Bearer with your wv_… key from the dashboard.
• Chat: POST chat/completions/ — OpenAI-compatible; optional stream: true; tools/function calling on DeepSeek & OpenRouter — Agent guide
• Voice & media: tts/, stt/, translation/, image/ — same credit rules as the web UI.
• Cursor / MCP: Local stdio MCP server (webvoice-mcp): agent registration, chat, TTS, STT, optional Solana top-up — Add to Cursor

One API for (almost) everything

With API keys tied to your account, you can programmatically access the same capabilities you see in the UI: synthesise speech, transcribe audio, translate text, generate images, send transactional email where enabled, check credit balance, and more. Responses are JSON-friendly (including Base64 payloads for audio and images), so integration stays straightforward in Node, Python, Swift, Kotlin, PHP, or low-code automations.

CORS is configured so browser-based front-ends on other domains can call the documented API routes safely, while server-to-server integrations use the same endpoints with your secret key.

Where it fits

• Mobile apps: Let users dictate notes, hear AI read content, or transcribe meetings without building models yourself.
• Websites & PWAs: Add voice accessibility, podcast previews, or image generation beside your existing pages.
• Back-office & automation: Batch TTS for notifications, STT for support tickets, translation for multilingual queues.
• Partner products: White-label the capabilities: your brand, our engines, one shared credit wallet.

Chat models in the catalogue

The same chat models you pick in the web UI are available via API forward. Per-request billing uses credits_per_request — models marked Free cost 0 and work without debiting your wallet.

Free via API (OpenRouter, 0 credits): openrouter/openrouter/free, openai/gpt-oss-20b:free, openai/gpt-oss-120b:free, google/gemma-4-31b-it:free — prefix with openrouter: in chat/completions. Groq-hosted models in the table below are billed normally. Limits (OpenRouter-aligned): 20/min; 50/day (1000/day after a credit purchase).

Safeguard-oriented models

Some entries are safeguard-class weights (for example OpenAI GPT-OSS variants tuned for stronger refusals and policy alignment). They are ideal for customer-facing or regulated workflows. Technical background, throughput context, and retention posture are covered in our security article. Secure AI models & Groq

Credits per request may differ from the generic defaults (e.g. Groq vs other hosts) when the catalogue assigns a custom value. The in-app model picker always shows the current debit before you send.

Other services (credits)

Beyond chat, integrations use the same wallet. Indicative rules on this deployment:

• Text-to-speech: 1 credits per 10-minute block (see pricing page for detail).
• Speech-to-text: 1 credits per 10-minute block (minimums may apply for very short audio).
• Translation: 0.1 credits per 1000 characters (minimum bill may apply).
• Image generation (MiniMax): 7 credits per image.
• Image generation (WAN 2.7 via WaveSpeed): 8 credits per image.
• Image generation (Nucleus via WaveSpeed): 4 credits per image (up to 2 images per request).
• Image generation (Runway Gen4 Image): 2 credits @ 720p, 3 @ 1080p (up to 3 reference images).
• Image generation (Runway Gen4 Image Turbo): 1 credits per image (2.5× faster, up to 3 reference images).
• Image generation (Stable Diffusion 3): 1 credits per image (typography, img2img optional).
• Image generation (Stability AI SDXL): 1 credits per image (high quality, img2img optional).
• Image generation (FLUX.2 Klein 9B LoRA): 1 credits per image (custom LoRA adapters).
• Image generation (Qwen-Image-Max via DeepInfra): 20 credits per image (upstream $0.075/image).
• Image-to-image (FLUX.2 Flash Edit): 1 credits per edit (1–4 input images).
• Image-to-video: Seedance 2.0 Turbo — 15 credits per 5s @ 720p, 16 @ 1080p (4–15s). Runway Gen4 Turbo — 2 credits per 5s (2–10s).
• Text-to-video (Seedance 2.0 Fast Turbo): 13 credits per 5s @ 720p, 14 @ 1080p (×2 with reference images/videos).
• Text-to-video (Google Veo 3.1 Lite): 7 credits per 6s @ 720p, 11 @ 1080p.
• Text-to-video (Google Veo 3.1 Fast via DeepInfra): 3.5 credits per second ($0.15/s upstream).
• Chat defaults when no per-model override is set: 0.5 credits per request on Groq routes, 1 on other providers — superseded by the table above when a model defines its own credits.

Daily allowance at login: 10 free credits (no expiration on purchased balance). Full price list

Credits and security

API usage draws from the same credit balance as the web app, with transparent per-operation costs. Keys can be created and revoked from the dashboard; keep them secret on servers and use short-lived tokens in clients when you proxy requests through your backend.

Choosing a safeguard model for an integration does not replace your own legal review — it reduces risk of unsafe completions for typical consumer-style prompts and aligns with our stated provider and retention policies.

Next steps

Read the interactive API documentation, create a test key after sign-up, and prototype with curl or your stack of choice. When you are ready, wire the calls into production flows — no separate “integration tier” required beyond credits.

Summary

• Full product in the browser + full-feature REST API
• Embed in apps, sites, automations — same account, same credits
• Documented endpoints for voice, text, images, and account utilities
• Chat catalogue with per-model credits; safeguard models called out explicitly
• TTS, STT, translation, and image generation share one credit wallet with clear rules