> For the full page index, fetch https://echonova.ai/llms.txt

# Quickstart: your first speech file

From a new workspace to a downloaded audio file in six steps, one sentence each, with the page to read when a step needs more.

Ten minutes after signing up you can have a sentence spoken by an Arabic
voice and saved as a file on your computer. This page is the shortest route:
six steps, one sentence each, and a link to the page that explains the step
when you want more. Nothing here needs code, a payment method or anything
beyond the trial credits every new workspace starts with.

## What you need

- An account and its workspace; see
  [Sign up and create your workspace](/docs/get-started/sign-up-and-create-a-workspace).
- Credits on the workspace balance. A new workspace's trial has
  them; the sidebar shows the plan's credits used out of its total.
- A sentence to speak. Any text up to 5,000 characters; a short one is
  cheapest for a first try.

## Steps

### Step 1: Open Text to Speech

Sign in and choose **Text to Speech** under **Playground** in the sidebar —
or use the **Text to Speech** tab of the studio on **Home** for a one-off
sentence; see
[Generate speech](/docs/playground/generate-speech#step-1-open-text-to-speech).

### Step 2: Type a sentence

Click into the writing area and type or paste your text, watching the
counter under it stay under `5000`; see
[Generate speech](/docs/playground/generate-speech#step-2-type-or-paste-your-text).

### Step 3: Pick a voice

Keep the voice shown under **Selected voice**, or press **Change voice**,
filter by **Dialect**, play a preview, choose one and press **Select**; see
[Choosing a voice](/docs/playground/choosing-a-voice).

### Step 4: Generate

Press **Generate** at the bottom of the settings panel, wait while
it reads **Generating...**, and listen when the player appears and starts
playing; see
[Generate speech](/docs/playground/generate-speech#step-5-generate).

### Step 5: Download the file

Press **Download** in the player and a `.wav` file named `speech-` and the
generation's identifier lands in your browser's downloads folder; see
[Formats, quality and download](/docs/playground/formats-quality-download#download).

### Step 6: Find it again later

Open the **History** tab of the settings panel, or **Assets** in the
sidebar, to play or download the same generation again for free; see
[Text to Speech history](/docs/playground/tts-history).

## Get it right

A generation is charged on the whole text as soon as you press the button,
so hear the voice on one sentence before you paste the whole script.

**Recommended**

- One sentence, listen, change the voice or the dialect if it is wrong, then
  the full text once.

**Less effective**

- The full 5,000 characters first, and a second full generation when the
  dialect turns out wrong.

## Worked example

Sara signs in, chooses **Text to Speech** under **Playground**, and pastes:

```text
أهلاً بكم في إيكونوڤا
```

She presses **Change voice**, sets **Dialect** to **Egyptian**, plays two
previews, picks one and presses **Select**. She presses **Regenerate
audio**; a few seconds later the player appears and the sentence plays.
She presses **Download** and `speech-` followed by an identifier and
`.wav` appears in her downloads folder. The button now shows what that
generation cost next to **Credits**, and the plan card in the sidebar has
moved by the same amount.

## What happens next

- The generation is in the **History** tab and on **Assets** for every
  member of the workspace, and replaying or downloading it again is free.
- The credits were taken from the workspace balance before the audio was
  made, and would have been returned had the generation failed. See
  [Text to Speech limits and credits](/docs/playground/tts-limits-and-credits).
- Back on **Home**, **Recents** now lists the sentence with its voice and
  dialect. Selecting that row opens it in the Text to Speech page and plays
  it for free; use **Open Assets** to do the same from **Assets**. See
  [A tour of the Home page](/docs/get-started/dashboard-tour#recents).

## Limits

| Limit       | Value                                      | Applies to      |
| ----------- | ------------------------------------------ | --------------- |
| Text length | 1 to 5,000 characters, counted with spaces | Each generation |
| Voices      | One voice per generation                   | Each generation |
| Output      | One WAV file per generation                | Each generation |

Every limit on the page, with the message each one shows, is on
[Text to Speech limits and credits](/docs/playground/tts-limits-and-credits).

## Troubleshooting

### The Generate button is greyed out

The writing area holds only whitespace, the voice list is still loading, or
a generation is already running. Type some text, or wait a moment.

### Error: Insufficient subscription balance

The workspace balance cannot cover the charge; nothing was charged or
generated. Add credits under **Subscription & billing**, or shorten the
text. See [Credits, top-ups and rollover](/docs/account/credits-top-ups-rollover).

### The audio plays but Download gives me a file I cannot open

The file is WAV. A tool that refuses it usually expects MP3; convert the
WAV with a tool of your choice. See
[Formats, quality and download](/docs/playground/formats-quality-download#can-i-get-mp3-from-the-portal).

## FAQ

### What will my first generation cost?

The charge counts the characters of your text without spaces, at the rate
on [Credits](/docs/developers/credits), with no rounding up. The
figure on the button before your first generation is not a quote for your
text; the figure after it is what it cost.

### Can I get MP3?

Every portal generation is WAV. Download it and convert it. See
[Formats, quality and download](/docs/playground/formats-quality-download#can-i-get-mp3-from-the-portal).

### The text is Arabic without diacritics. Should I add them?

With a **Fusha** voice selected, switch **Tashkeel** on before you
generate; it is free, and the marks fix how each word is pronounced. See
[Add diacritics](/docs/playground/add-diacritics).

### Can I do the same from code?

Yes. The same generation is one API call; see
[Text to Speech quickstart](/docs/developers/tts-quickstart).

## Related

- [Generate speech](/docs/playground/generate-speech): every control on the
  Text to Speech page
- [Choosing a voice](/docs/playground/choosing-a-voice): what to listen for
  before you commit to a voice
- [Voice settings](/docs/playground/voice-settings): the sliders to leave
  alone on a first try, and what they do when you are ready
- [Text to Speech limits and credits](/docs/playground/tts-limits-and-credits):
  the limits and the charge
- [Quickstart: your first agent](/docs/get-started/quickstart-first-agent):
  give that voice a conversation