August 28, 2026AIExplore

How to Design a Voice with ElevenLabs

Write structured ElevenLabs Voice Design prompts, pair matching preview text, generate previews, and save a production ready voice when the library falls short.

ElevenLabs

Voice Design generates new voices from a text description when the Voice Library does not already fit. Official Voice Design docs recommend a structured prompt: Native language, gender, age, quality, then persona, emotion, and timbre or pacing detail. Each generate produces preview options and charges based on preview text characters. Docs also note Professional Voice Clones remain the highest consistency option when a library PVC fits. Overview: /explore/elevenlabs.

This guide shows how to write design prompts, pick preview text that matches the persona, and save keepers. Pair with remixing at /blog/how-to-use-elevenlabs-voice-remixing and narration at /blog/how-to-use-elevenlabs-for-text-to-speech-narration.

When to design instead of browse

Design when you need a specific age, accent, or character the library lacks. Browse first. Official guidance prefers an existing PVC when one already matches production needs.

Step by step Voice Design workflow

1. Write the structured prompt

Start with Native language and dialect. Add gender, age, quality, persona, emotion, then one or two sentences on timbre and pacing. Avoid FX words like reverb unless intentional.

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Design skeleton Native English, neutral American accent. Female, 35 to 40. Studio quality. Persona: trusted product narrator. Emotion: calm, confident, warm. Smooth timbre, relaxed pacing, clear emphasis on key terms.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Design skeleton Native English, neutral American accent. Female, 35 to 40. Studio quality. Persona: trusted product narrator. Emotion: calm, confident, warm. Smooth timbre, relaxed pacing, clear emphasis on key terms." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Inputs:
Use the actual source material required for this task, together with the intended audience, destination, format, and quality requirements. Specify relevant files, text, URLs, structured data, reference assets, dimensions, language, tone, timing, model or generation settings, credentials, field mappings, or integration values when they apply. Keep secrets out of the example and use the product's supported credential or configuration mechanism.

Workflow:
1. Start with one representative input that contains the important characteristics of the real workload.
2. Configure elevenlabs specifically for the requested task and make the important settings explicit.
3. Run the first pass and inspect the result against the objective.
4. Correct only the settings or source material responsible for a failed requirement instead of changing everything at once.
5. Validate the intermediate result before passing it to the next step.
6. When the result meets the requirements, save the successful configuration so the same process can be repeated consistently.
7. If this is a multi-step workflow, verify each handoff and keep a clear fallback or review path for failures.

Requirements:
The example must use supported functionality only. Do not invent product features, pricing, limits, integrations, model names, API behavior, or unavailable settings. Preserve important source data and formatting. Validate required fields before processing, handle empty or invalid input explicitly, prevent accidental duplicate processing where relevant, and stop or route the item for review when a required step fails. For credentials, use the supported secure configuration rather than placing secrets directly in the content.

Expected output:
The final result should directly satisfy the "Design skeleton Native English, neutral American accent. Female, 35 to 40. Studio quality. Persona: trusted product narrator. Emotion: calm, confident, warm. Smooth timbre, relaxed pacing, clear emphasis on key terms." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

2. Pair matching preview text

Official docs warn that mismatched preview text fights the prompt. Calm voices need calm sample lines.

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Preview match Prompt: calm reflective Good preview: It has been quiet lately. I have had time to think. Bad preview: Hey! I cannot stand what you have done!!!. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Preview match Prompt: calm reflective Good preview: It has been quiet lately. I have had time to think. Bad preview: Hey! I cannot stand what you have done!!!" that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Preview match Prompt: calm reflective Good preview: It has been quiet lately. I have had time to think. Bad preview: Hey! I cannot stand what you have done!!!" requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

3. Generate previews and save the best

Compare three options when offered. Save the keeper, then run real narration tests before locking a campaign.

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Save checklist [ ] Accent correct [ ] Age feel correct [ ] No unwanted FX [ ] Works on Multilingual v2 or v3 as needed. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Save checklist [ ] Accent correct [ ] Age feel correct [ ] No unwanted FX [ ] Works on Multilingual v2 or v3 as needed" that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Save checklist [ ] Accent correct [ ] Age feel correct [ ] No unwanted FX [ ] Works on Multilingual v2 or v3 as needed" requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Practical Voice Design examples

Product narrator

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native English, neutral American. Female, 35 to 40. Studio quality. Persona: trusted product narrator. Emotion: calm, confident.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native English, neutral American. Female, 35 to 40. Studio quality. Persona: trusted product narrator. Emotion: calm, confident." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native English, neutral American. Female, 35 to 40. Studio quality. Persona: trusted product narrator. Emotion: calm, confident." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Airline agent

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native French, français standard. Female, late 20s. Excellent quality. Persona: airline agent. Emotion: reassuring, precise.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native French, français standard. Female, late 20s. Excellent quality. Persona: airline agent. Emotion: reassuring, precise." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native French, français standard. Female, late 20s. Excellent quality. Persona: airline agent. Emotion: reassuring, precise." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Fantasy guide

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native English. Male, 40s. Excellent quality. Persona: thoughtful quest guide. Emotion: curious, gentle.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native English. Male, 40s. Excellent quality. Persona: thoughtful quest guide. Emotion: curious, gentle." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native English. Male, 40s. Excellent quality. Persona: thoughtful quest guide. Emotion: curious, gentle." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Newsreader

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native English, received pronunciation leaning. Female, 30s. Studio quality. Persona: clear newsreader. Emotion: neutral, precise.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native English, received pronunciation leaning. Female, 30s. Studio quality. Persona: clear newsreader. Emotion: neutral, precise." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native English, received pronunciation leaning. Female, 30s. Studio quality. Persona: clear newsreader. Emotion: neutral, precise." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Tired New Yorker humor

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native English, New York. Female, 60s. Ok quality. Persona: dry humor neighbor. Emotion: wry, warm.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native English, New York. Female, 60s. Ok quality. Persona: dry humor neighbor. Emotion: wry, warm." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native English, New York. Female, 60s. Ok quality. Persona: dry humor neighbor. Emotion: wry, warm." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Kids coach careful

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native English. Male, early 30s. Studio quality. Persona: patient coach. Emotion: encouraging, clear.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native English. Male, early 30s. Studio quality. Persona: patient coach. Emotion: encouraging, clear." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native English. Male, early 30s. Studio quality. Persona: patient coach. Emotion: encouraging, clear." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Trailer tease

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Dramatic cinematic narrator, deep male, controlled intensity. Preview with epic but grammatical sentences.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Dramatic cinematic narrator, deep male, controlled intensity. Preview with epic but grammatical sentences." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Dramatic cinematic narrator, deep male, controlled intensity. Preview with epic but grammatical sentences." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Support Gulf Arabic

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native Arabic, soft Gulf influence. Female, 30 to 40. Excellent quality. Persona: customer service. Emotion: warm, polite.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native Arabic, soft Gulf influence. Female, 30 to 40. Excellent quality. Persona: customer service. Emotion: warm, polite." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native Arabic, soft Gulf influence. Female, 30 to 40. Excellent quality. Persona: customer service. Emotion: warm, polite." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Polish storyteller

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Native Polish. Male, 48 to 58. Excellent quality. Persona: creative dreamer. Emotion: curious, inviting.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Native Polish. Male, 48 to 58. Excellent quality. Persona: creative dreamer. Emotion: curious, inviting." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Native Polish. Male, 48 to 58. Excellent quality. Persona: creative dreamer. Emotion: curious, inviting." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Simple calm male

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: A calm male narrator. Short prompt when you want a neutral default.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "A calm male narrator. Short prompt when you want a neutral default." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "A calm male narrator. Short prompt when you want a neutral default." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

More design prompts

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Prompt: Native Spanish, europeo. Female, 35 to 40. Ok quality. Persona: support operator. Emotion: reassuring.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Prompt: Native Spanish, europeo. Female, 35 to 40. Ok quality. Persona: support operator. Emotion: reassuring." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Prompt: Native Spanish, europeo. Female, 35 to 40. Ok quality. Persona: support operator. Emotion: reassuring." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Prompt: Native Japanese. Female, 20s. Excellent quality. Persona: soft assistant. Emotion: polite, bright.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Prompt: Native Japanese. Female, 20s. Excellent quality. Persona: soft assistant. Emotion: polite, bright." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Prompt: Native Japanese. Female, 20s. Excellent quality. Persona: soft assistant. Emotion: polite, bright." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Prompt: Native German. Male, 50s. Studio quality. Persona: documentary host. Emotion: measured.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Prompt: Native German. Male, 50s. Studio quality. Persona: documentary host. Emotion: measured." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Prompt: Native German. Male, 50s. Studio quality. Persona: documentary host. Emotion: measured." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Prompt: Avoid FX: no phone filter, no reverb, no tape hiss.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Prompt: Avoid FX: no phone filter, no reverb, no tape hiss." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Prompt: Avoid FX: no phone filter, no reverb, no tape hiss." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic elevenlabs workflow is being prepared for Use ElevenLabs for Text to Speech Narration using elevenlabs. The starting requirement is: Prompt: Guidance: raise adherence when accent accuracy matters most.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Prompt: Guidance: raise adherence when accent accuracy matters most." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Prompt: Guidance: raise adherence when accent accuracy matters most." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Tips

Lead with Native language
Match preview text to persona
Avoid accidental FX words
Browse library PVCs before designing
Test saved voices on real scripts
Watch preview character credit use

Keep a short generation log: voice id or name, model id, Stability, Similarity, and the one change you will try next. That habit turns two regenerations into learning instead of noise.

When credits, plan features, or model access differ from memory, open elevenlabs.io/pricing and the official models page. Limits change, so do not promise client deliverables from outdated screenshots.

Treat the first take as a draft. Listen for clipped words, wrong stress on names, and accidental stage direction read aloud before you rewrite the whole script.

If you collaborate, store winning scripts and voice settings on a shared page. Reuse the skeleton and change only the line that failed.

For product work, decide latency versus expressiveness before you pick a model. Flash favors speed. Multilingual v2 favors long form stability. v3 favors performance tags.

Separate exploration from release. Explore voices freely, then re check commercial usage rules on a paid plan before monetizing outputs.

Close each session by naming keeper takes and discarding near misses. A clean library saves time when a producer asks for the calm narrator from Tuesday.

Prefer clear punctuation and short sentences over vague vibe words alone. The model reads what you write.