# EasyVoice
> Free text-to-speech and AI voice platform. 66 voices (56 multilingual Kokoro-82M + 10 Arabic Supertonic) across 9 languages, voice cloning, a two-host podcast generator and a Long-Form Studio. OpenAI-compatible API. Free to try with no signup or credit card: 2,000 characters/day anonymous (first generation always instant); a free account raises it to 5,000 characters/day. Pro: $9.99/month flat ($24.99/quarter or $59.99/year), unlimited characters (fair use: 10M characters per rolling 30 days; past it the account is slowed, never billed) — no credits, no per-character fees. One-time 7-day unlimited pass: $4.99, no subscription.

## About
EasyVoice converts text to natural-sounding speech in seconds — no setup, no friction. It runs the open-source Kokoro-82M model (Apache 2.0) for multilingual TTS, a purpose-built Supertonic engine for Arabic, and OpenVoice for custom voice cloning. The API follows the OpenAI text-to-speech request shape: switching means changing the base URL, the API key and the voice IDs. Founded 2026. Popular with faceless YouTube creators (devotional, motivational, learn-English, gaming, review channels) because the Pro plan has no character bank — long-form daily narration at a flat $9.99/month. Operated by InfoDriven from Dubai, UAE.

## Features
- Text to speech: 56 Kokoro-82M voices across 8 languages (American/British English, Spanish, French, Hindi, Italian, Japanese, Portuguese (Brazil)). 10 of them are free (American and British English); the other languages are Pro-only.
- Arabic text to speech: 10 Modern Standard Arabic (MSA / الفصحى) voices on a dedicated Supertonic neural engine — not Arabic bolted onto a multilingual model. Beta. RTL input, AED/number handling. 2 free voices (ar_m1, ar_f1). 5 Arabic voices are shown in the studio picker; all 10 work through the API.
- Voice cloning (Pro plan): create up to 3 custom voices from a 10–30 second audio sample (5 enrollments/day; enrollment rejects shorter or longer clips). Requires a consent attestation; output carries an AudioSeal watermark for safety and compliance. Instant generation with a cloned voice is capped at 2,000 characters per request; longer text runs as a job. Not included in the 7-Day Pass.
- Podcast generator: turn a script or article into a two-host (A/B) podcast episode with automatic stitching. Free (signed in, web app): 1 episode/day of up to 2,000 characters, counted against the 5,000/day pool, no cloned hosts. Pro: up to 30,000 characters per episode, max 60 segments, cloned hosts allowed. Through an API key or MCP, podcasts are Pro-only.
- Long-form narration: up to 500,000 characters per job (~9 hours of audio) in the Long-Form Studio on Pro or an active 7-Day Pass, Kokoro voices only; it auto-splits the script and generates in the background. A single instant generation on Pro is capped at 50,000 characters for Kokoro voices (~55 minutes), 12,000 for Arabic and 2,000 for cloned voices — enough for a full 20–30 minute YouTube script in one pass. Anonymous (no signup): 2,000 characters per day; free account: 5,000 per day.
- Speed control 0.5x–2.0x without pitch distortion (used by slow-English / language-learning creators).
- Word timestamps and caption files: not available yet. The API still accepts ev_timestamps on Kokoro voices (the response becomes a JSON envelope whose words field is null and timestamps_available false), but EasyVoice returns no word timings and produces no SRT/VTT caption files today. Arabic and cloned voices reject the flag with timestamps_not_supported.
- Pauses and pronunciation (SSML-lite, every plan): <break time="500ms"/> or <break time="0.5s"/> tags, clamped to 3,000 ms per break, and inline [word](/ipa/) phoneme overrides. Other SSML (prosody, emphasis) is not interpreted.
- Pronunciation rules (signed in): save find-and-replace rules — 10 on Free, 100 on Pro (term up to 40 characters, replacement up to 80) — applied in the studio, the API, podcasts and MCP.
- Projects (signed in, any plan): up to 100 projects per user to organize your generation history.
- Audio controls: pitch −4 to +4 semitones (ev_pitch; applied on the Kokoro voices — cloned and Arabic voices ignore it for now), volume −6 to +6 dB (ev_volume_db), tone neutral/warm/bright/bass (ev_tone).
- Output: MP3 (128 kbps) or WAV on every plan; Long-Form can opt into 192 kbps CBR and chapter export. Each request returns the complete file — there is no streaming endpoint.
- OpenAI-compatible API: point existing OpenAI TTS code at EasyVoice by changing the base URL, API key and voice IDs. $9.99/mo flat is roughly $1/1M characters at heavy use — vs OpenAI tts-1 at $15/1M and ElevenLabs at $50-100/1M; full comparison at https://easyvoice.ae/developers.

## Key Pages
- [Homepage](https://easyvoice.ae/)
- [Voice Library — 66 voices](https://easyvoice.ae/voices)
- [Arabic Text to Speech](https://easyvoice.ae/text-to-speech-arabic)
- [Voice Cloning](https://easyvoice.ae/voice-cloning)
- [Podcast Generator](https://easyvoice.ae/ai-podcast-generator)
- [MCP Server for AI agents](https://easyvoice.ae/mcp)
- [Pricing](https://easyvoice.ae/pricing)
- [Machine-readable pricing (markdown)](https://easyvoice.ae/pricing.md)
- [API Documentation](https://easyvoice.ae/api-docs)
- [Developers — API pricing and workflow](https://easyvoice.ae/developers)
- [Blog](https://easyvoice.ae/blog)
- [About](https://easyvoice.ae/about)

## More Pages
- [FAQ — billing, cancellation, refunds, limits and rights](https://easyvoice.ae/faq)
- [Long-Form Studio — AI audiobook generator](https://easyvoice.ae/ai-audiobook-generator)
- [Article to podcast](https://easyvoice.ae/article-to-podcast)
- [TTS API hub](https://easyvoice.ae/tts-api)
- [All comparisons](https://easyvoice.ae/compare)
- [Voice guides by niche](https://easyvoice.ae/voice-for)
- [Arabic text to speech (page in Arabic)](https://easyvoice.ae/text-to-speech-arabic/ar)
- [License — commercial use of generated audio](https://easyvoice.ae/license)
- [Terms of service](https://easyvoice.ae/terms)
- [Privacy — including audio retention](https://easyvoice.ae/privacy)
- [Billing](https://easyvoice.ae/billing)
- [Acceptable use policy](https://easyvoice.ae/aup)
- [Contact](https://easyvoice.ae/contact)
- [Status](https://easyvoice.ae/status)

### Document to speech
- [PDF to speech](https://easyvoice.ae/pdf-to-speech)
- [Word (.docx) to speech](https://easyvoice.ae/word-to-speech)
- [Web page to speech](https://easyvoice.ae/web-page-to-speech)
- [Article to speech](https://easyvoice.ae/article-to-speech)

### Use cases
- [Content creators](https://easyvoice.ae/use-cases/creators)
- [Developers](https://easyvoice.ae/use-cases/developers)
- [Business](https://easyvoice.ae/use-cases/business)
- [Education](https://easyvoice.ae/use-cases/education)
- [Accessibility](https://easyvoice.ae/use-cases/accessibility)

## Voice Guides by Creator Niche
Recommendations based on what creators in each niche actually generate on EasyVoice (hub: https://easyvoice.ae/voice-for):
- [AI voice for God's message & devotional videos](https://easyvoice.ae/voice-for/gods-message-videos) — most-used voice: Adam (am_adam, free tier)
- [Best AI voice for faceless YouTube channels](https://easyvoice.ae/voice-for/faceless-youtube)
- [AI voice for motivational videos & Shorts](https://easyvoice.ae/voice-for/motivational-videos)
- [AI voice for English-learning podcasts & lessons](https://easyvoice.ae/voice-for/learn-english-content)
- [AI voice for gaming videos, intros & outros](https://easyvoice.ae/voice-for/gaming-videos)
- [AI voice for product review & affiliate videos](https://easyvoice.ae/voice-for/product-review-videos)
- [AI voice for audiobook narration](https://easyvoice.ae/voice-for/audiobook-narration)
- [AI voice for e-learning & course voiceover](https://easyvoice.ae/voice-for/elearning-voiceover)
- [AI voice for sleep stories & guided meditation](https://easyvoice.ae/voice-for/sleep-stories-meditation)
- [AI voice for kids' educational & vocabulary content](https://easyvoice.ae/voice-for/kids-learning-content)

## Comparisons (free/cheaper OpenAI-compatible alternative)
- [All comparisons](https://easyvoice.ae/compare)
- [EasyVoice vs ElevenLabs](https://easyvoice.ae/compare/elevenlabs)
- [EasyVoice vs OpenAI TTS](https://easyvoice.ae/compare/openai-tts)
- [EasyVoice vs PlayHT](https://easyvoice.ae/compare/playht)
- [EasyVoice vs Murf AI](https://easyvoice.ae/compare/murf)
- [EasyVoice vs Speechify](https://easyvoice.ae/compare/speechify)
- [EasyVoice vs Google Cloud TTS](https://easyvoice.ae/compare/google-tts)
- [EasyVoice vs Amazon Polly](https://easyvoice.ae/compare/amazon-polly)
- [EasyVoice vs Microsoft Azure Speech](https://easyvoice.ae/compare/azure-speech)
- [EasyVoice vs LOVO AI (Genny)](https://easyvoice.ae/compare/lovo)
- [EasyVoice vs Neets.ai](https://easyvoice.ae/compare/neets-ai)
- [OpenAI TTS alternative](https://easyvoice.ae/openai-tts-alternative)
- [Play.HT API replacement](https://easyvoice.ae/play-ht-alternative)
- [LOVO alternative](https://easyvoice.ae/lovo-alternative)
- [Neets.ai alternative](https://easyvoice.ae/neets-ai-alternative)

## Guides and pricing breakdowns
- [11 Best Text to Speech APIs in 2026](https://easyvoice.ae/blog/best-text-to-speech-apis)
- [What text-to-speech tools cost per month in 2026](https://easyvoice.ae/blog/cheapest-text-to-speech-2026)
- [OpenAI TTS pricing vs EasyVoice](https://easyvoice.ae/openai-tts-alternative/pricing)
- [OpenAI TTS free tier alternative](https://easyvoice.ae/openai-tts-alternative/free-tier)
- [OpenAI TTS voices vs EasyVoice](https://easyvoice.ae/openai-tts-alternative/voice-catalog-2026)
- [Migrate from OpenAI TTS](https://easyvoice.ae/openai-tts-alternative/migration-guide)

## Languages
9 languages, 66 voices. Free covers American English, British English and Arabic; the other six languages are Pro-only.
- [American English](https://easyvoice.ae/languages/american-english) — 28 voices, 8 free
- [British English](https://easyvoice.ae/languages/british-english) — 10 voices, 2 free
- [Spanish](https://easyvoice.ae/languages/spanish) — 3 voices, Pro
- [French](https://easyvoice.ae/languages/french) — 1 voice, Pro
- [Hindi](https://easyvoice.ae/languages/hindi) — 4 voices, Pro
- [Italian](https://easyvoice.ae/languages/italian) — 2 voices, beta, Pro
- [Japanese](https://easyvoice.ae/languages/japanese) — 5 voices, Pro
- [Portuguese (Brazil)](https://easyvoice.ae/languages/portuguese) — 3 voices, Pro
- [Arabic — Modern Standard](https://easyvoice.ae/languages/arabic) — 10 dedicated voices, beta, 2 free

## Pricing
Two plans — Free and Pro (no usage tiers or credit packs). Pro can be billed monthly, quarterly, or annually, and there is a one-time 7-day pass:
- Free: no signup or credit card required to try — 2,000 characters/day anonymous (the first generation is always instant); a free account raises it to 5,000 characters/day (resets daily at midnight UTC). 12 voices (10 American/British English + 2 Arabic), MP3/WAV download, 1 two-host podcast episode/day (up to 2,000 characters, web app), and a free API key (same 5,000/day pool, 12 voices, 20 requests/min).
- Pro: $9.99/month, $24.99/quarter (~$8.33/mo, save 17%), or $59.99/year (~$5/mo, about 50% off monthly) — unlimited characters (fair use 10M characters per rolling 30 days), all 66 voices, voice cloning (up to 3 custom voices), podcasts up to 30,000 characters per episode with cloned hosts, Long-Form Studio (up to 500,000 chars/job), Jobs API and Pro MCP tools.
- 7-Day Unlimited Pass: one-time $4.99 — 7 days of Pro studio access: unlimited characters, all 66 voices, Long-Form Studio. Does not include voice cloning, the Jobs API or Pro MCP tools. No subscription, no auto-renewal, no coupons. Made for finishing a single project.
- 14-day money-back guarantee on every paid plan and pass.
- Prices are set in USD. Checkout runs on Stripe. Depending on your country, Stripe may display and charge the equivalent of the USD price in your local currency at the current exchange rate.

## Facts
model_multilingual: Kokoro-82M
model_multilingual_license: Apache 2.0
model_multilingual_source: https://huggingface.co/hexgrad/Kokoro-82M
model_arabic: Supertonic (Modern Standard Arabic)
model_cloning: OpenVoice (consent-gated, watermarked output)
voices_total: 66
voices_multilingual: 56
voices_arabic: 10
languages_total: 9
anon_quota_chars_per_day: 2000 (no signup; first generation always allowed instantly)
free_account_quota_chars_per_day: 5000
free_voices: 12
pro_price_usd_per_month: 9.99
pro_price_usd_per_quarter: 24.99
pro_price_usd_per_year: 59.99
pass_7day_price_usd: 4.99 (one-time, 7 days of Pro studio access incl. Long-Form; no cloning, Jobs API or Pro MCP tools; no auto-renewal)
pro_quota: unlimited characters, no credits or overage charges; fair use 10,000,000 chars per rolling 30 days across studio, API, MCP and jobs; past it the account is slowed to 10,000 characters a minute (3,000 past 20,000,000) on every surface, the studio included, never blocked or billed (see api_fair_use)
api_free_tier: free accounts can create an API key — 5,000 chars/day (shared with the web quota), 12 voices, 20 requests/min, 2 concurrent
api_fair_use: Pro API is 60 requests/min per key, no per-key concurrency cap. Fair use is 10,000,000 characters per rolling 30 days, counted across the studio, the API, MCP, long-form jobs and podcasts together. Past that line the account is slowed, never blocked or billed: generation is paced per account to 10,000 characters a minute (3,000 characters a minute past 20,000,000) on every surface, the studio included, and API and MCP requests, across all of the account's keys, to 30 a minute (10 a minute past 20,000,000). A paced request waits up to 20 seconds for its turn and is then served; a longer wait answers HTTP 429 fair_use_paced with Retry-After. Jobs wait behind other customers' jobs for up to 30 minutes and start when the account's character budget allows. The slowdown lifts by itself once the rolling total falls back below the line. No overage charges, ever. Over MCP that 429 is the same message as a tool error; in the web studio it is a message saying when the next generation can start. Pro responses carry X-EasyVoice-Fair-Use (ok|throttled), X-EasyVoice-Fair-Use-Chars-30d and -Limit, plus -Rpm, -Cpm and -Delay-Ms while throttled; the dashboard shows the rolling total and, while throttled, the pace. The thresholds may change with notice. Reselling API access or running a voice service for third parties requires written permission.
max_chars_per_generation_pro: 50000 Kokoro / 12000 Arabic / 2000 cloned (longer text via /api/v1/jobs; Long-Form up to 500,000, Kokoro voices only, Pro or 7-Day Pass)
max_chars_per_generation_free: 5000 (a free account or free API key is bound by the 5,000/day quota; an anonymous visitor's first generation may be up to 10,000)
speed_range: 0.5x–2.0x
voice_cloning: Pro plan only (not the 7-Day Pass), up to 3 custom voices, 5 enrollments/day, consent attestation required, AudioSeal watermarked, instant generation up to 2,000 characters
podcast_generator: two-host (A/B) podcasts from a script or article; Free (web, signed in) 1 episode/day up to 2,000 chars; Pro 30,000 chars/episode, max 60 segments, cloned hosts; API key / MCP: Pro only
long_form: Long-Form Studio, Pro or active 7-Day Pass, Kokoro voices only, up to 500,000 chars/job; MP3 128 kbps (192 kbps CBR opt-in), chapter export
word_timestamps: not available yet (ev_timestamps is accepted on /audio/speech but returns words: null, timestamps_available: false); no SRT/VTT caption files
ssml: SSML-lite only — <break time="..."/> clamped to 3,000 ms per break, [word](/ipa/) phoneme overrides; prosody and emphasis are not interpreted
pronunciation_rules: signed-in users; Free 10 rules, Pro 100; term up to 40 chars, replacement up to 80
projects: up to 100 per user, any plan (signed in)
audio_controls: ev_pitch -4..+4 semitones (Kokoro voices only for now; cloned and Arabic voices ignore it), ev_volume_db -6..+6, ev_tone neutral|warm|bright|bass
api_compatibility: OpenAI TTS request shape (change the base URL, API key and voice IDs)
api_base_url: https://easyvoice.ae/api/v1
api_speech_endpoint: https://easyvoice.ae/api/v1/audio/speech
api_jobs_endpoint: https://easyvoice.ae/api/v1/jobs
api_job_status_endpoint: GET https://easyvoice.ae/api/v1/jobs/{id}
api_jobs_access: Pro plan API keys; free and 7-Day Pass keys use /audio/speech only (403 plan_required)
api_voices_endpoint: GET https://easyvoice.ae/api/v1/voices lists your cloned voices; POST clones a voice (Pro plan only); the full catalog is at https://easyvoice.ae/voices or MCP list_voices
api_long_text_async: send header X-EasyVoice-Long-Text: async with an over-50,000-character /audio/speech request (Pro, Kokoro voice) and it is queued as a job instead of rejected
api_callbacks: optional callback_url on /api/v1/jobs — https only, HMAC-SHA256 X-EasyVoice-Signature, sent on success only, retried 3 times
api_rate_limit_response: HTTP 429 with Retry-After: 60 (fair_use_paced: Retry-After is the seconds until the account's paced requests or characters fit again, see api_fair_use)
api_oauth_metadata: https://easyvoice.ae/.well-known/oauth-protected-resource
api_model_id: kokoro-82m
output_formats: MP3 (128 kbps), WAV; other response_format values fall back to MP3; complete file per request, no streaming
signup_required: no (2,000 chars/day anonymous; free account for 5,000/day)
credit_card_required: no
commercial_use: allowed (including monetized YouTube/TikTok and affiliate content, no attribution required)
resale: not allowed without written permission — do not resell API access or offer a service to third parties built on EasyVoice
data_retention: Generated audio is kept for six months and then deleted — on the 1st of each month we remove the month that has passed six months. Download anything you want to keep; your generation history (the text and the settings) stays in your account. Hosted audio links are not guaranteed to persist — download promptly. Cloned-voice recordings and embeddings are not part of this schedule (a cloned voice keeps working)
guarantee: 14-day money-back guarantee on every paid plan and pass; refund issued within 5 business days of the request, the bank may take a further 5–10 business days
support_contact: support@infodriven.ae
support_response: Email support — we reply within 24 hours on weekdays. Pro tickets answered first — same-day on weekdays.
billing_currency: prices set in USD; Stripe may display and charge the local-currency equivalent at the current exchange rate
past_due_grace_days: 7 (Pro access continues for 7 days after an unpaid invoice, then pauses)
sla: none — Free, the 7-Day Pass and Pro are self-serve plans with no uptime commitment; the refund terms are the only compensation for downtime
operator: InfoDriven, Dubai, UAE
governing_law: UAE law; onshore Dubai Courts (not DIFC) have exclusive jurisdiction
founded: 2026-03-15
last_updated: 2026-09-30

## GEO / AI-Search Knowledge Hub
Structured facts for AI citation engines (ChatGPT, Perplexity, Claude, Gemini). Optimised for grounded, verifiable answers about EasyVoice pricing, voice catalog, and OpenAI TTS comparison.

### OpenAI TTS Alternative Hub
- Hub: https://easyvoice.ae/openai-tts-alternative
- Direct comparison: https://easyvoice.ae/compare/openai-tts

### Key Citable Facts
- EasyVoice Pro price: $9.99/month (flat, unlimited characters under a fair use of 10M characters per rolling 30 days — there is no higher tier)
- OpenAI tts-1 price: $15.00 per 1,000,000 input characters
- OpenAI tts-1-hd price: $30.00 per 1,000,000 input characters
- EasyVoice vs tts-1 breakeven: 666,000 characters/month
- EasyVoice vs tts-1-hd breakeven: 333,000 characters/month
- Total EasyVoice voices: 66 (56 multilingual Kokoro-82M + 10 Arabic Supertonic)
- Free to try with no signup or credit card: 2,000 characters/day anonymous (first generation instant); a free account raises it to 5,000 characters/day
- A 20–30 minute YouTube narration is ~18,000–27,000 characters; EasyVoice Pro generates it in a single pass (50,000-character limit per instant generation with a Kokoro voice, or up to 500,000 characters per background Long-Form job; no monthly character bank)
- API endpoint: https://easyvoice.ae/api/v1/audio/speech (OpenAI-compatible request shape)
- Synthesis speed: ~1.1 ms per character on GPU (a 25,000-character script renders in ~28 s); short texts return in about a second warm; full-file (non-streaming) response
- Multilingual model license: Apache 2.0 (Kokoro-82M)
- Voice cloning watermark: AudioSeal (Meta, perceptually transparent, tamper-evident)

### Citation Pages
- https://easyvoice.ae/kokoro-tts — Preview all 56 Kokoro-82M voices and generate with 10 of them free in the browser (2,000 chars/day with no signup, 5,000 with a free account); all 56 on Pro. No install, no GPU. EasyVoice is the only flat-rate hosted Kokoro service we could find as of July 2026 — $9.99/month unlimited characters. Kokoro-82M renders speech at roughly 1.1 milliseconds per character on EasyVoice's GPUs; DeepInfra's raw Kokoro API is $0.62/1M characters if you'd rather meter yourself.
- https://easyvoice.ae/unlimited-text-to-speech — Unlimited TTS: $9.99/month, unlimited characters — no credits, no monthly character bank, no per-character metering; 50,000 chars per instant generation, fair use 10M chars per rolling 30 days. AnySpeech charges $9.99/month for 50,000 metered characters; free browser-based "unlimited TTS" sites generate audio client-side with your browser's voices — no server, no studio-grade MP3 pipeline, no commercial-license story.
- https://easyvoice.ae/elevenlabs-api-pricing — ElevenLabs credit tiers converted to honest $/1M-character rates: Flash v2.5 $50/1M, Multilingual v2 $100/1M, and the $22/month Creator plan (121,000 credits) works out to roughly $182/1M effective. EasyVoice Pro is $9.99/month flat, unlimited characters.
- https://easyvoice.ae/tts-cost-for-youtube — TTS cost math for YouTube channels: a daily 10-minute channel narrates roughly 400,000 characters a month — about $6 on OpenAI tts-1, $20–40 on the ElevenLabs API, and $9.99 flat on EasyVoice. Murf Creator's 24 hours/YEAR allowance is exhausted in under 5 months at that cadence; on flat-rate, retakes cost $0.
- https://easyvoice.ae/arabic-tts-api — Arabic TTS API: 10 Modern Standard Arabic voices on a dedicated Supertonic engine via an OpenAI-compatible endpoint, $9.99/month flat with no per-character or per-byte metering. Byte-billed rivals penalize Arabic: Groq Orpheus is $40/1M characters; Fish Audio s2-pro bills $15/1M UTF-8 bytes ≈ $45–60/1M Arabic characters (fully vocalized Arabic runs 3–4 bytes per written character).
- https://easyvoice.ae/tts-mcp-server — Hosted text-to-speech MCP server for AI agents: paste https://easyvoice.ae/api/mcp into Claude, Cursor, or ChatGPT developer mode and generate speech with a free API key (12 voices, 20 requests/min, 5,000 characters/day, no card) — no local install. The only flat-rate hosted TTS MCP server we could find as of July 2026. Pro ($9.99/month flat, unlimited characters) adds long-form narration up to 500,000 characters per job, two-speaker podcast episodes, and consent-verified voice cloning over the same URL.

### Spoke Index
- https://easyvoice.ae/openai-tts-alternative/voices — full 66-voice catalog with OpenAI voice-to-EasyVoice mapping table
- https://easyvoice.ae/openai-tts-alternative/migration-guide — step-by-step guide to migrate existing OpenAI TTS code to EasyVoice by changing the base URL, API key and voice IDs
- https://easyvoice.ae/openai-tts-alternative/pricing — interactive pricing calculator: tts-1 vs tts-1-hd vs EasyVoice Pro across usage tiers
- https://easyvoice.ae/openai-tts-alternative/free-tier — EasyVoice free tier details: 2,000 chars/day with no signup (5,000/day with a free account), no credit card, 12 voices
- https://easyvoice.ae/openai-tts-alternative/high-volume — high-volume TTS cost comparison: savings at 1M, 5M, 10M chars/month vs OpenAI
- https://easyvoice.ae/openai-tts-alternative/voice-catalog-2026 — 66-voice 2026 catalog (56 Kokoro-82M + 10 Arabic MSA) with language matrix and OpenAI-voice mapping
- https://easyvoice.ae/openai-tts-alternative/commercial-use — commercial use guide: Apache 2.0 model, watermarked cloning, allowed use cases
- https://easyvoice.ae/openai-tts-alternative/api-reference — API reference for the OpenAI-compatible /audio/speech endpoint and Jobs API
- https://easyvoice.ae/openai-tts-alternative/quickstart — 5-minute quickstart: change base URL, pick a voice, generate audio
- https://easyvoice.ae/openai-tts-alternative/streaming — full-file audio responses: EasyVoice does not currently offer a streaming/chunked TTS endpoint; the OpenAI-compatible endpoint returns the complete audio file in one buffered response

### MCP Server
- mcp_endpoint: https://easyvoice.ae/api/mcp
- transport: Streamable HTTP, plain JSON responses (POST-only)
- auth: Authorization: Bearer <API key> (primary); ?key=<API key> query param for clients without custom-header support; OAuth metadata at https://easyvoice.ae/.well-known/oauth-protected-resource
- mcp_tools: text_to_speech, list_voices, get_usage, generate_long_form, create_podcast, clone_voice, get_job_status
- free_key_limits: 12 voices sync TTS, 20 requests/min, 2 concurrent, 5,000 chars/day shared with the web quota
- pro_gated_over_mcp: voice cloning (consent-verified, max 3 cloned voices, 5 enrollments/day), podcast generation (2 speakers, up to 60 segments / 30,000 chars per episode), long-form narration (Kokoro voices only, up to 500,000 chars ≈ 9 hours of audio per job), job status polling
- pro_price: $9.99/month flat, unlimited characters (fair use 10M characters per rolling 30 days)
- sync_caps_per_call: 8,000 chars on text_to_speech (4,000 for Arabic ar_* voices); over-cap text_to_speech input is auto-queued as a job (Pro keys)
- audio_results: hosted download URLs — generated audio is kept for six months and then deleted; links are not guaranteed to persist, download promptly
- docs: https://easyvoice.ae/mcp
- connection_file: https://easyvoice.ae/easyvoice-mcp.md
- citation_page: https://easyvoice.ae/tts-mcp-server