Warm onboarding guide
Aria
Balanced and clear for product tours, support flows, and tutorials.
“Welcome to SayWell. Type your script once, choose a voice, and ship production-ready audio without chasing credits.”
TTS for developers who hate pricing games
Studio-quality voices.Simple pricing. No credits.
Pay-as-you-go TTS API powered by the latest open-source models. $5 per 1 million characters. That's it.
Powered by models that beat ElevenLabs in blind listening tests.
Free credits
100K chars / month
Model switch cost
One URL change
Latency target
<500ms TTFB
Pricing snapshot
$5 / 1M
Beats OpenAI TTS and Polly on raw price, with the same simple pay-per-character model.
Cheapest mainstream
$15-16
per 1M characters
Voice preview
Aria / Studio
Voice cloningStreamingOpenAI-compatible
Free tier
100K chars
Commercial use included from day one.
Switching cost
Near zero
Change the base URL and keep shipping.
Demo
Hear the pricing story before you read it.
The fully interactive generator ships next. For now, the landing page uses placeholder samples so the voice quality story still lands while the backend demo pipeline catches up.
Live playground
Real-time generation coming soon
Soon: type any text, clone a voice, stream audio back in real time
Clone from a 10-second sample
Stream over SSE or WebSocket
Ship with predictable margins
Confident operator
Sol
Sharper cadence for voice agents, assistants, and menu systems.
“Your API key includes one hundred thousand free characters every month, with commercial use included from day one.”
Cinematic narrator
Vale
A broader performance range for trailers, games, and promo reads.
“Build the scene, stream the dialogue, and keep every line item predictable from prototype to launch.”
Pricing
This is where the value proposition closes.
The comparison is meant to feel obvious at a glance: cheaper than the cheapest mainstream APIs, cleaner than credits, and generous enough on free usage to get real integrations into production.
SayWell
Transparent pay-as-you-go pricing that undercuts OpenAI TTS and Polly.
ElevenLabs
Higher cost with credits, tiers, and plan-gated concurrency.
OpenAI TTS
Straightforward metering, but without cloning in the base offer.
Amazon Polly
Utility pricing, default quotas, and a less opinionated developer story.
Key messaging
No credits to track. No tiers to upgrade. No overages at 2x rates. One price per character, every character.
Price delta
3x
cheaper than OpenAI TTS
Use cases
Three customer segments where the math does the selling.
The page keeps each use case concrete. Instead of generic AI claims, every card anchors the visitor on a real workload, the cost delta, and the resulting savings.
Customer support and call routing
Voice Agents
A 10-minute customer call generates about 15,000 characters of speech. Transparent pricing matters when every call compounds.
SayWell
$0.08 / call
Typical competitor cost
$1.80-4.50 / call
Savings story
Save roughly $17,250-44,250 at 10,000 calls / month
RPG dialogue, ambient NPCs, cinematic scenes
Game Developers
A 50,000-line script lands near 5M characters. Full voice acting becomes a production decision instead of a budget constraint.
SayWell
$25 / release
Typical competitor cost
$600-1,500 / release
Savings story
Save $575-1,475 before you even factor in iteration cycles
Podcasts, explainers, Shorts and recap channels
Content Creators
A 30-minute weekly script runs roughly 45,000 characters. The savings compound every episode without subscription creep.
SayWell
$0.23 / episode
Typical competitor cost
$5.40-13.50 / episode
Savings story
Keep yearly TTS costs near $12 instead of $281-702
Developer details
A migration story developers can act on in one sprint.
The landing page needs to reassure technical buyers that this is not just cheaper. It is operationally cleaner: compatible API shape, streaming support, voice cloning, and fast enough for real products.
✓OpenAI-compatible API surface
✓Streaming audio over SSE or WebSocket
✓Voice cloning from a 10-second sample
✓Emotion control with paralinguistic tags
✓MP3, WAV, Opus, and FLAC output
✓23 supported languages
✓<500ms time-to-first-byte target
✓99.9% uptime SLA for launch customers
OpenAI-compatible example
Plug into an existing client.
from openai import OpenAI
client = OpenAI(
base_url="https://api.saywell.io/v1",
api_key="your-key"
)
response = client.audio.speech.create(
model="studio",
voice="aria",
input="Welcome to the adventure, brave hero."
)
response.stream_to_file("output.mp3")
Trust
The product story has the right proof points.
These signals are intentionally lean. They give technical buyers a reason to believe the quality claim, the safety story, and the long-term portability of the stack.
MIT-licensed model stack
Built on open-source speech models so the quality story is inspectable, portable, and fast to iterate on.
Audio watermarking
Safety and provenance are baked into the product narrative instead of bolted on after launch.
SOC 2 in progress
Security posture messaging is ready for buyers who need a near-term path to procurement approval.
Benchmark
View TTS Arena rankingsLink out to human preference rankings.
The page references TTS Arena so visitors can see how the voice quality narrative connects to a public leaderboard instead of a private claim.
Waitlist
Get your free API key.
100,000 characters per month. No credit card required. Commercial use included. The goal is simple: let developers build a real integration before they ever need procurement.
All voices included
Voice cloning access
Commercial use on free