> For the full page index, fetch https://echonova.ai/llms.txt # Browse the voice library Open the voice library, read what a voice row tells you, hear a preview, send a voice to Text to Speech and put it on an agent. The voice library is where every voice you can use starts: the voices you generate speech with under **Text to Speech**, and the voices an agent speaks with. By the end of this page you will have opened the library, read a voice row, heard a preview, sent a voice to Text to Speech and put a voice on an agent. ## What you need - An EchoNova account and a workspace. See [What is EchoNova](/docs/get-started/what-is-echonova) if you are new. - To add a voice to an agent: an agent that is not archived. See [Create an agent](/docs/agents/create-an-agent). ## Steps ### Step 1: Open the library Select **Voices** in the sidebar. The page opens on **Every voice has a hometown**, with a line above it that reads **Library**, the number of voices and the number of ways of speaking. Two tabs sit at the top: **Library** and **My voices**. **Library** opens first, with a search box that reads **Search a name, a tone or a use**, a **Filters** button, a **Sort** menu and a row of dialect chips beneath it. A button at the top right, **Add a voice**, is described under [What happens next](#what-happens-next). Below the chips, **Library** has up to three sections: - **Most used** (kicker **Popular**): up to six voices that have been used, most used first, each with a play control. It is hidden while no voice matches your filters. - **Every voice, by where it comes from**: the voice rows, grouped by dialect. A dialect with no voices yet shows a line that reads, for example, **Egyptian · coming soon**. - **Voices for what you build**: one button per use, such as **Contact centre** or **Narration and audiobooks**, listing its voices. Press one to filter the library to that use. ### Step 2: Read a voice row Every row on **Library** and under **My voices** has the same shape: | Part | What it shows | | ------------------------- | ------------------------------------------------------------------------------------- | | The orb | Plays and pauses the voice's preview clip | | The name | The voice's display name, for example **Ahmed**. Press it to open the voice's details | | The meta line | The dialect, gender and age, for example **Egyptian · Male · Middle** | | The description | One sentence about the voice, then its tone, pace and uses | | **Use in Text to Speech** | Opens Text to Speech with this voice selected | | The **+** button | **Add to My voices**; it reads **In My voices** once added | | A play control | Plays and pauses the preview clip | Press anywhere else on a row to open the voice's details panel. It shows the voice's tags, **Used by** (the number of people) and **Voice id**, with **Play**, **Add to My voices**, **Use in Text to Speech**, **Add to an agent**, **Copy id** and **Similar voices**. Rows under **My voices** replace the **+** button and the play control with a **More** menu. It holds **Use in Text to Speech**, **Add to an agent**, **Copy id** and **Remove from My voices**. **Add to an agent** shows only if your role can manage agents. Every workspace sees the same library. A **Most used** card shows the name, the tone and pace, the dialect and one use, with a play control and the **+** button. ### Step 3: Preview a voice Press the orb or the play control on a row or on a **Most used** card. The player opens at the bottom of the page and the clip starts. Press the orb again to pause. One clip plays at a time. Starting a second voice replaces the first. Leaving the page stops it. The player's controls are described in [Filters, previews and favourites](/docs/playground/filters-previews-favourites). ### Step 4: Browse one dialect Press a dialect chip. The voice rows show only that dialect, and the chip stays pressed until you press it again. Each chip shows how many voices the dialect has. A dialect with no voices yet appears after the chips with the word **soon** and cannot be pressed. See [Dialects](/docs/playground/dialects) for what each dialect means. ### Step 5: Send a voice to Text to Speech Press **Use in Text to Speech** on the row or in the voice's details panel. **Text to Speech** opens with that voice selected and its dialect set to match, so the next thing you do is type and generate. See [Generate speech](/docs/playground/generate-speech). ### Step 6: Add a voice to an agent 1. Press **Add to an agent** in the voice's details panel, or in the **More** menu of a row under **My voices**. You see it only if your role can manage agents. 2. A dialog titled **Add voice to agent** opens, with the line **Select an agent to add this voice to.** Type in **Search agents** to narrow the list. 3. Every active agent in the workspace is listed by name, with its description under it, or **Voice agent** when it has none. Archived agents are not listed. 4. Select an agent and press **Add voice**. The button reads **Adding…** while it saves. Press **Cancel** or the close control to leave without adding. A confirmation reads **Voice added** with **The voice was added to Acme Support.**, naming the agent you chose. If the voice is already on that agent, nothing changes and the same confirmation shows. ## Worked example You are building a support agent for Cairo and want to hear Egyptian voices first. Open **Voices**, press the **Egyptian** chip, and the rows show only Egyptian voices. Press each orb in turn and let the clips play in the player. You settle on one. Press its name to open the details panel, then **Add to an agent**, type `support` in **Search agents**, select **Acme Support** and press **Add voice**. The confirmation reads **The voice was added to Acme Support.** Then press **Use in Text to Speech** on the same voice. **Text to Speech** opens with the voice selected, so you can generate a greeting for the agent's welcome message with the same voice. ## What happens next The **Add a voice** button opens a sheet titled **Three ways in**. **Clone from a recording** takes you to voice cloning. **Design from a description** and **Remix a library voice** are marked **Coming soon**. Every voice you add is private to your workspace. A voice you add to an agent appears on that agent's **Voice** tab under **Used Voices**, on the default speech settings, and it is not the default voice. Open the agent to make it the default or to tune it; see [Voices and language](/docs/agents/voice-and-language). The agent can use the new voice from its next conversation. A voice you send to **Text to Speech** is the selected voice for what you generate next. Generating speech is charged in credits; browsing and previewing are not. ## Limits | Limit | Value | Applies to | | --------------------------------------- | ----------------- | ----------------------------- | | Voices the library shows | 100 | **Library** and **My voices** | | Voices in **Most used** | 6 | **Library** | | Agents listed in **Add voice to agent** | 100 active agents | The dialog | ## Troubleshooting ### We couldn't load the voices. The library did not load. Press **Try again** and, if the message returns, reload the page. ### No agents found. No active agent matches what you typed in **Search agents**, or the workspace has no active agent. Clear the search. Archived agents are not offered; restore the agent or create one, then add the voice. ### This voice was not added. Try again. The save did not go through. Press **Add voice** again. To confirm the result, open the agent's **Voice** tab and look for the voice under **Used Voices**. ### The orb does nothing That voice has no preview clip, so nothing plays. Choose another voice, or press **Use in Text to Speech** and generate a sentence on **Text to Speech** to hear it on your own text. ## FAQ ### Does previewing a voice cost credits? No. Previews are fixed sample clips and are free to play. Credits are charged when you generate speech or when an agent speaks. ### Can I add my own voice to the library? You can add a voice of your own with **Add a voice**, then **Clone from a recording**. It is private to your workspace and appears under **My voices**, not in the shared library. ### Is the library different in each of my workspaces? No. Every workspace sees the same library. What differs per workspace is which voices its agents carry. ### Does adding a voice to an agent change what the agent says? No. The voice sets how the agent sounds. What it says comes from its persona and prompt; see [Write the persona](/docs/agents/persona). ### Where do I see the voice's id for the API? Open a voice's details panel, or the **More** menu of a row under **My voices**, and press **Copy id**. The panel also shows the id under **Voice id**. You can also fetch the voices call in the Developer API; see [Text to Speech voices and options](/docs/developers/tts-voices-and-options). ## Related - [Filters, previews and favourites](/docs/playground/filters-previews-favourites): narrow the list, use the player and keep a shortlist - [Dialects](/docs/playground/dialects): what each dialect means and how to pick one for a region - [Choosing a voice](/docs/playground/choosing-a-voice): narration versus agents, and what to listen for - [Generate speech](/docs/playground/generate-speech): what happens after **Use in Text to Speech** - [Voices and language](/docs/agents/voice-and-language): the agent's **Voice** tab, the default voice and per-voice tuning - [Text to Speech voices and options](/docs/developers/tts-voices-and-options): the same voices from the API