Expressive text-to-speech

Great text-to-speech for LLM apps, videos, games, and products

Create polished voice output in minutes. Write normal scripts, add natural-language delivery cues in brackets, reuse the same voice across every prompt, and let your LLM app generate audio through MCP.

Start $2.99 test See setup guide

Natural expressionSave voicesConnect LLM apps

USER

Write a warm dashboard update.

ASSISTANT

[warm] Welcome back. [excited] Report ready.

Audio previewaudio ready

Voice: Product Guide v2
Expression: [calm and bright], [urgent but friendly]
Use for: dashboard updates
Output: audio ready

For agent platformsGame creatorsYouTube narrationSupport productsEducation apps

What is TextToSpeechSkills?

TextToSpeechSkills is a paid text-to-speech platform for people who want good voice output without setting up a complicated audio stack. You can write a normal script, add full natural-language expression directions like [trying not to wake someone] or [confident but playful], save a voice template, and generate audio from the studio, API, MCP server, or installed skills. The product is built for LLM workflows, so a non-technical user can ask an agent to prepare narration while the account owner keeps billing, keys, templates, and permissions controlled. Developers get async speech jobs, webhooks, scoped API keys, workspace billing, and predictable credits. One credit is one full minute of audio, which makes testing and production usage easier to explain before a team connects speech to apps, videos, games, courses, support, onboarding, or internal tools without rebuilding the content model later.

Easy workflow

From text to polished audio without fiddly voice prompts

The workflow is simple enough for a creator and structured enough for a product team. Start in the studio, then reuse the same natural expression markup and saved voices from your app or LLM workflow.

Write or paste text

Use a script, app message, or LLM draft.

Direct the performance

Use plain-language cues like [quiet], [trying not to wake someone], or [loud and angry].

Pick a saved voice

Templates keep every prompt consistent.

Generate audio

Preview in the UI or hand it to your app.

Why teams choose it

Easy to start, good enough to keep using

Make great speech quickly, keep voices consistent, and let humans or LLM apps use the same simple setup.

Sounds intentional

Natural-language expression cues make emotion, pacing, and emphasis clear, so speech feels directed instead of guessed.

Stays consistent

Voice templates keep your narrator, character, support voice, or course instructor recognizable across many prompts.

Easy for LLM users

Connect once, choose the voices your app may use, and ask your LLM workflow to create audio from approved text.

Ready for teams

Workspaces, scoped keys, and clear credit plans help you move from one test to shared production use.

Popular workflows

Voice workflows for games, videos, agents, and learning

Start in the UI, connect an LLM app through MCP, or move the same workflow into the API when it becomes part of your product.

Create voices for your own game

Prototype NPC dialogue, mission updates, tutorials, and character barks with saved voice templates.

Narrate YouTube videos

Turn scripts, outlines, and LLM drafts into consistent narration for tutorials, explainers, and Shorts.

Give AI agents voice output

Let agents turn approved scripts into audio without teaching them a custom API first.

Support voice agents

Keep spoken help replies clear, on brand, and tied to approved support voices.

Course and training narration

Create repeatable instructor voices for lessons, internal training, onboarding, and accessibility audio.

Product demos and explainers

Generate polished narration for demos, release notes, feature tours, and product education.

LLM setup

Follow a short setup guide, choose a saved voice, and ask your LLM app to create audio.

Read setup guide

Easy LLM setup

Give your LLM app a voice tool without building an integration first

TextToSpeechSkills is made for the moment when you want an LLM to help with speech, not just write scripts. Install the MCP tool, choose which voices are allowed, and let the same workflow create narration, character lines, support replies, or product audio.

Read the guide

MCP install

Paste one command into your LLM app settings.

Saved voices

Pick the narrator, character, or support voice it can use.

Natural expression

Ask for natural directions like [quiet], [excited but restrained], or [loud and angry].

Audio ready

Review the result in your dashboard or send it back to your app.

Credit pricing

Start with a small paid test, then upgrade when usage is real

Every plan includes the UI, API, MCP setup, natural expression markup, and saved voice templates. One credit is one full minute of audio.

Test

$2.99 / month

Yearly: $29.99 / year

30 credits = 30 full minutes included

Try UI, API, and MCP
Create saved voice templates
Upgrade when you need more credits

Monthly Yearly

Starter

$12 / month

Yearly: $120 / year

300 credits = 300 full minutes / month

300 full minutes of audio
API, UI, and MCP included
Good for small projects

Monthly Yearly

Pro

Best fit

$39 / month

Yearly: $390 / year

1,200 credits = 1,200 full minutes / month

1,200 full minutes of audio
Workspace add-on available
Best for regular content and apps

Monthly Yearly

Scale

$99 / month

Yearly: $999 / year

3,200 credits = 3,200 full minutes / month

3,200 full minutes of audio
Workspace add-on available
Built for production usage

Monthly Yearly

Workspaces are available on Pro and higher for $2 per user per month with central billing.

Expression markup

Natural-language voice direction that stays readable

Write expressive delivery notes directly in brackets. Examples like [quiet] are starters; you can use full phrases such as [trying not to wake someone] or [excited but professional].

Starter	Category	Example	How to extend it
quiet	volume	`[quiet] hello there.`	Speak softly or at a low volume. You can combine it with natural language when the scene needs more detail.
whisper	volume	`[whisper] I have a secret.`	Use a very soft, intimate delivery. You can combine it with natural language when the scene needs more detail.
loud	volume	`[loud] Listen up!`	Increase emphasis and volume. You can combine it with natural language when the scene needs more detail.
excited	emotion	`[excited] that's amazing!`	Speak with high energy and enthusiasm. You can combine it with natural language when the scene needs more detail.
angry	emotion	`[angry] how could you do that?`	Speak with anger or frustration. You can combine it with natural language when the scene needs more detail.
warm	tone	`[warm] welcome back.`	Use a friendly and pleasant tone. You can combine it with natural language when the scene needs more detail.
serious	tone	`[serious] this is important.`	Use a focused, earnest tone. You can combine it with natural language when the scene needs more detail.
fast	pace	`[fast] let's go!`	Increase speaking speed. You can combine it with natural language when the scene needs more detail.
slow	pace	`[slow] take your time.`	Decrease speaking speed. You can combine it with natural language when the scene needs more detail.
pause	timing	`[pause] wait here.`	Insert a short pause. You can combine it with natural language when the scene needs more detail.
laugh	interjection	`[laugh] that's funny.`	Add a brief natural laugh. You can combine it with natural language when the scene needs more detail.

Reusable voices

Templates keep one voice across many prompts

Save persona, tone, pacing, accent, style rules, and sample prompts as versioned assets your whole team can reuse.

Template	Persona	Baseline voice	Direction
Calm Narrator	Clear narrator for explainers and product walkthroughs	Charon	Calm, reassuring, precise. Keep emotional changes controlled unless inline tags request otherwise.
Energetic Coach	High-energy coach for launches and motivating clips	Puck	Energetic and encouraging without sounding frantic.
Warm Teacher	Friendly teacher for lessons and onboarding	Sulafat	Friendly, patient, and clear. The listener should feel safe asking the next question.

Template preview

Calm Narrator v1

approved

Compare saved versions across multiple prompts before making a template active.