Frequently asked questions
Which voice is best for TikTok?
There is no single best voice; it depends on your content style. Use Aria for clear narration, Nova for upbeat energy, or Felix for energetic promos. You can test all eight presets via the /v1/voices endpoint.
What is the difference between Aria and Nova?
Aria is described as a warm, clear narrator voice, ideal for straightforward explanations. Nova is bright and upbeat, suited for energetic content. Both support all 23 languages and can be adjusted with speed settings between 0.25 and 4.0 to match your video’s pacing requirements.
Can I use Felix for educational content?
Yes. Felix is an energetic, promo-style male voice. While designed for promos, you can adjust the speed parameter (0.25–4.0) to make it suitable for educational content. There are no voice restrictions based on content type, only on character limits and language support.
Is Orion too deep for shorts?
Orion is a deep, documentary-style male voice. Whether it is too deep depends on your creative preference. You can increase the speed parameter up to 4.0 to make the delivery faster and less drawn out, ensuring it fits well within short-form video constraints.
Do voices sound robotic?
Voices are based on licensed recordings from LibriTTS-R and LibriVox. Each generated file carries an inaudible watermark marking it as synthetic speech. While natural, they are distinct from human recordings. They do not use custom voice cloning, so they sound like the eight preset voices.
Can I mix voices in one video?
The API generates one audio file per request. To mix voices, you must send separate requests for each text segment using different voice parameters. You then join the resulting audio files on the client side. There is no built-in voice mixing in a single API call.
Which voice sounds most like Siri?
There is no official comparison to Siri. Atlas is a neutral, news-style voice. Aria is a clear narrator. You can test all eight presets via the API to determine which best matches your desired vocal characteristic for your project.
Are the voices natural sounding?
Voices are based on licensed recordings from LibriTTS-R and LibriVox. Each file includes an inaudible watermark indicating synthetic origin. They are not custom-cloned human voices. Quality depends on the preset chosen and speed settings. They are designed for clear, programmatic voiceover generation rather than perfect human mimicry.
Can I use Luna for storytelling?
Yes. Luna is explicitly described as a soft, storytelling voice. It supports all 23 languages. You can adjust the speed between 0.25 and 4.0 to control the pacing of your story, ensuring it matches the visual rhythm of your TikTok, Reels, or Shorts video.
Is Milo good for conversational tone?
Yes. Milo is described as a friendly, conversational male voice. It is well-suited for direct-to-camera or dialogue-style content. You can fine-tune the delivery speed using the speed parameter to ensure the tone remains natural and engaging for your audience without requiring manual recording.
What makes Atlas neutral?
Atlas is categorized as a neutral, news-style voice. It lacks the specific emotional tags of other presets like Aria’s warmth or Nova’s brightness. This neutrality makes it versatile for factual content. It supports all 23 languages and can be sped up or slowed down as needed.
Do voices support emojis in text?
The API accepts text input up to 4096 characters. Emojis are part of the text input and will be processed. Emojis are processed as text input. The language parameter helps ensure correct pronunciation.
Can I change voice mid-video?
No, a single API request generates one audio file with one voice. To change voices mid-video, you must send multiple requests with different voice parameters. You then concatenate the resulting audio files in your video editing software. There is no multi-voice support within a single response.
Are there age restrictions for voices?
There are no explicit age restrictions for using the voices. The voices are presets based on licensed recordings. You can use any voice for any content type. The only limits are technical: 4096 characters per request, 60 requests per minute, and valid language codes.
Female voices overview
The female voices in our API offer distinct tonal qualities tailored for short-form video content. Each preset is designed to convey specific tonal qualities, allowing creators to match the audio to the visual narrative without manual recording.
Aria: warm narrator
Aria provides a warm, clear narration style ideal for storytelling or educational content. Her voice cuts through background music effectively, ensuring the message remains intelligible.
- Best for: Narrative-driven shorts and tutorials
Nova: bright and upbeat
Nova delivers a bright, energetic tone perfect for fast-paced edits. Her upbeat delivery keeps viewer attention high, making her suitable for lifestyle or entertainment clips.
- Best for: Trending audio and quick cuts
Luna: soft storytelling
Luna offers a soft, gentle voice ideal for emotional or calming content. Her tone works well for voiceovers that require a soothing or intimate atmosphere.
- Best for: Aesthetic vlogs and emotional hooks
Male voices overview
The male preset collection offers three distinct vocal profiles tailored for different short-form content styles. Each voice is designed for clarity and consistency, ensuring your voiceovers remain professional whether you are promoting a product or delivering news.
Felix: energetic promo
Felix delivers high-energy narration ideal for promotional clips. His tone is upbeat and dynamic, making him suitable for quick cuts and fast-paced product highlights.
- Ideal for promotional clips and ads
- High-energy, upbeat delivery
- Best for quick cuts
Milo: friendly conversational
Milo provides a warm, approachable sound perfect for conversational content. His style feels natural and engaging, suitable for storytelling or direct-to-camera scripts.
- Warm, approachable tone
- Natural and engaging delivery
- Suits conversational scripts
Atlas: neutral news
Atlas offers a clear, neutral voice optimized for informational content. His consistent pace and lack of bias make him ideal for news-style updates or educational snippets.
- Clear, neutral tone
- Consistent pacing
- Optimized for news/updates
Deep voice option
Orion delivers a deep, documentary-style narration that adds gravitas to short-form content. Orion’s deep, documentary-style voice adds gravitas to short-form content. This makes it ideal for serious topics or high-stakes reveals where tone matters more than speed.
The voice remains clear even at faster speeds, ensuring listeners catch every detail. Its steady cadence works well for storytelling segments, historical context, or product reviews that require a trustworthy, authoritative presence. Creators often pair Orion with cinematic music to elevate production value.
Orion: documentary style
Orion uses a deep, resonant tone optimized for narrative depth. The voice maintains clarity at various speeds, supporting rates from 0.25 to 4.0. It avoids the ‘cheerful’ pitch of Nova or the ‘instructional’ calm of Sage, leaning instead into a neutral, serious delivery.
This preset is built for consistency. Whether narrating a 15-second hook or a 60-second explainer, Orion’s voice does not waver in tone. It provides a stable audio bed that allows visuals and music to remain the primary focus, rather than competing with a high-energy vocal.
- Deep, resonant male voice
- Optimized for narrative depth
- Maintains clarity at high speeds
- Neutral, serious delivery
Use cases for deep voice
Deep voices like Orion excel in faceless channels that focus on facts, history, or tech reviews. The lower register signals authority, which helps retain viewers during educational segments. It also works well for ‘storytime’ videos where the creator wants to sound like a narrator rather than a friend.
Use Orion for product reviews that need a ‘premium’ feel. The voice adds a layer of professionalism to unboxing clips or tech demos. It is also effective for horror, mystery, or true-crime style content where a lighter voice might feel out of place.
- Faceless TikTok channels
- Tech and product reviews
- Storytime and narrative clips
- Horror or mystery content
Tone comparison
Compared to Felix (energetic promo) or Milo (friendly conversational), Orion feels less personal and more observational. Felix is best for hype and quick sales pitches; Orion is better for context and depth. Milo invites the viewer in; Orion stands back and lets the story unfold.
Luna and Aria offer warmth and clarity, which is great for tutorials. Orion, however, adds weight. If your content is about to make a serious point or reveal a twist, Orion’s lower pitch creates a subtle shift in mood that signals importance to the viewer.
- Less personal than Milo
- More observational than Felix
- Adds weight vs. warmth of Luna/Aria
- Signals importance or twists
Choosing based on content type
Selecting the right voice ensures your short-form content resonates with its intended format. Whether you are explaining a concept, entertaining viewers, or selling a product, the voice must match the pacing and tone of the visual narrative to maintain engagement without manual recording.
The API allows precise control over speed and voice selection, enabling you to tailor the audio output specifically for educational clarity, high-energy entertainment, or direct promotional calls to action.
Educational content
Use Luna for soft storytelling or Sage for calm, instructional narration. These voices work well for faceless explainer videos where clarity is key. Set the speed to 1.0 or slightly lower to ensure every detail is heard clearly without rushing the listener.
Educational clips benefit from consistent pacing. The 4096-character limit per request allows you to generate explanations in single calls.
- Use Sage for instructional clarity
- Use Luna for storytelling
- Keep speed at 1.0 for clarity
Entertainment videos
Nova and Felix bring bright, upbeat energy suitable for viral-style content. Nova’s clear tone works well for quick cuts, while Felix adds an energetic promo feel that keeps viewers engaged during fast-paced editing sequences.
For entertainment, speed can be increased up to 2.0 or 3.0 to match rapid visual transitions. Requests are processed one after another, so under load a request may wait in a queue.
- Use Nova for upbeat energy
- Use Felix for energetic promos
- Increase speed for fast cuts
Promotional clips
Atlas provides a neutral, news-style delivery ideal for straightforward product reviews or announcements. His tone is reliable and professional, ensuring the message is delivered without distracting emotional inflections.
Promotional content often requires clear calls to action. Use Milo for a friendly, conversational approach if the brand voice is more personal, or stick to Atlas for a clean, direct presentation of features and pricing.
- Use Atlas for neutral announcements
- Use Milo for friendly tone
- Ideal for product reviews
Voice and audience match
Matching voice personas to your target demographic helps align audio tone with viewer expectations. The API’s eight distinct presets allow you to align audio tone with viewer expectations without manual recording.
Selecting the right voice ensures your message resonates correctly, whether you are targeting Gen Z on TikTok or professionals on LinkedIn.
Younger demographics
For fast-paced TikTok and Reels content targeting younger audiences, high-energy voices perform best. Nova (bright, upbeat) and Felix (energetic) capture attention quickly with their lively cadence.
- Nova for upbeat, energetic content
- Felix for promotional or hype clips
Professional audiences
When addressing professionals, clarity and neutrality are key. Atlas (neutral, news) provides a trustworthy, journalistic tone suitable for educational or corporate shorts.
- Atlas for neutral, news-style delivery
- Sage for calm, instructional explanations
General viewers
For broad appeal, warm and clear voices work universally. Aria (warm, clear narrator) and Milo (friendly, conversational) offer approachable tones that engage diverse audiences without niche constraints.
- Aria for warm, clear narration
- Milo for friendly, conversational style
Speed and tone interaction
Adjusting speed alters how the synthetic voice interacts with your video’s pacing. The API supports rates from 0.25x to 4.0x, allowing you to sync audio tightly with visual cuts or slow down narration for clarity.
Different speeds change the perceived tone without changing the voice preset. A higher speed creates urgency, while a lower rate adds weight to each word, affecting how viewers perceive the content's energy.
Fast-paced editing
Increasing speed up to 2.0x or higher helps compress longer scripts into short-form videos. This is useful for rapid-fire facts or quick product highlights where every second counts.
Be cautious with extreme speeds above 3.0x, as clarity may drop. Test with your specific voice preset to ensure the audio remains intelligible when synced with fast visual transitions.
- Use 1.5x–2.0x for quick updates or news-style clips
- Avoid exceeding 4.0x unless the text is very short
- Sync visual cuts to speech beats for better retention
Calm narration
Setting speed between 0.75x and 0.9x creates a relaxed, deliberate tone suitable for storytelling or emotional content. This works well with voices like Luna or Milo, which already have a softer baseline.
Slower pacing gives viewers more time to process information. It is ideal for educational explainers or reflective shorts where the message matters more than speed.
Ensure the background music does not compete with the slowed voice. Lower speeds can make audio feel sluggish if not paired with appropriate visual pacing.
- Best for Luna (storytelling) or Sage (instructional)
- Ideal for emotional or reflective content
- Pair with minimal music to avoid clutter
Emphasis effects
While the API lacks SSML tags, you can simulate emphasis by adjusting speed locally. Slowing down specific words or phrases draws attention without needing complex formatting.
Use slight speed variations to highlight key points. For example, slow down to 0.8x for a crucial statistic, then return to normal speed for the rest of the sentence.
This technique works best with clear, distinct voices like Aria or Felix, which maintain clarity even at varied speeds.
- Slow down for key statistics or calls to action
- Use speed shifts to mark topic changes
- Combine with visual cues for maximum impact
Technical considerations
Short-form video production requires precise audio handling to ensure voiceovers remain clear against background music and sound effects. Proper gain staging and EQ adjustments prevent the synthetic voice from being masked or distorted in the final edit.
Voice clarity in noise
Synthetic speech can sometimes clash with dense audio layers. Use high-pass filters on background tracks to preserve vocal presence, ensuring the AI voice remains distinct without requiring excessive volume boosts that introduce hiss.
- Use EQ to carve out space for vocal frequencies
- Avoid overlapping music in the 2kHz-4kHz range
- Monitor levels on mobile speakers for clarity checks
Compression artifacts
Heavy compression on the final mix can introduce pumping effects that highlight the synthetic nature of the voice. Apply gentle limiting rather than aggressive compression to maintain natural dynamics and prevent audio clipping during peak moments.
- Limit peak levels to -1dBTP
- Avoid sidechain compression on the voice track
- Check for inter-sample peaks in WAV exports
Format compatibility
Export your final voiceovers in MP3 or AAC for social platforms to minimize file size while maintaining quality. The API supports MP3, WAV, OPUS, FLAC, and PCM, allowing you to choose the optimal format for your editing workflow without re-encoding.
- MP3 is ideal for quick uploads
- WAV preserves maximum quality for editing
- OPUS is a supported format

Licensing and usage
ShortForm AI provides a reliable text-to-speech API built for creators making TikTok, Reels, and Shorts. We generate voiceovers programmatically, allowing you to produce content without manual recording.
Licensed recordings
Our eight voices are based on licensed recordings from the LibriTTS-R dataset (CC BY 4.0). These voices originate from LibriVox public-domain audiobook readings, ensuring a legally sound foundation for your generated audio files.
Commercial use allowed
You own the generated audio files and can use them commercially in your short-form videos. There are no additional licensing fees or attribution requirements for the underlying voice models, making this API suitable for monetized content.
Watermark presence
Every generated file carries an inaudible watermark that marks it as synthetic speech. This watermark is embedded directly into the audio data and does not affect the listening experience or the visual quality of your video.
Testing voices
Testing ensures the voice fits your video’s pacing before you commit to full production. Short-form creators need to verify clarity and tone quickly.
Use the included trial credit to generate samples without paying. Compare outputs directly against your script to catch mismatches early.
Use free trial
Sign up to receive $0.50 of credit, valid for 7 days. This covers approximately 50,000 characters, enough to test multiple voices.
Generate several short clips using different voices. Verify that the audio quality meets your standards before scaling up.
- Free trial credit: $0.50
- Valid for 7 days
- Covers ~50,000 characters
- No credit card required
Generate sample clips
Create 15–30 second audio clips for each voice option. Use the same text to ensure a fair comparison of tone and style.
Test voices with typical TikTok dialogue. Check for natural phrasing and clear articulation at your chosen speed settings.
Compare side by side
Play samples sequentially to identify the best fit for your specific niche. Note which voices maintain clarity during fast-paced edits.
Evaluate how each voice handles emphasis and pauses. Select the one that best matches your brand’s audio identity.
Multi-language voices
The API supports 23 languages, including English, Spanish, French, German, and Japanese. Any voice can speak any of these languages, allowing you to localize content without hiring new talent.
If the language is not specified, the system guesses from the script. Latin text defaults to English, so always set the language parameter for non-English content to ensure accurate pronunciation.
English voices
All eight voices handle English natively. Use Aria for clear tutorials or Nova for upbeat shorts. The engine processes text at roughly real-time speed, so plan your editing timeline accordingly.
- Native English pronunciation
- All 8 voices available
- Default language for Latin script
Other language support
Support includes Arabic, Chinese, Danish, Dutch, Finnish, Greek, Hebrew, Hindi, Italian, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Swahili, Swedish, and Turkish. Vietnamese and Thai are not supported.
Set the language field explicitly. This prevents the default English guess from mispronouncing non-English words in Latin scripts.
- 23 total languages
- Explicit language code recommended
- No support for Vietnamese, Thai, Ukrainian
Voice consistency
The same voice preset maintains consistent tone across different languages. This allows you to keep a unified brand identity even when producing multilingual content.
Audio files retain the same characteristics regardless of the spoken language, ensuring seamless transitions in your video edits.
- Consistent tone across languages
- Unified brand identity
- No voice cloning needed
Common use cases
Generate voiceovers programmatically for short-form content without manual recording. This API lets you scale production by sending text and receiving audio files, ideal for high-volume creators.
Use the API to maintain consistent audio quality across hundreds of videos, ensuring your audience hears the same clear tone regardless of the voice preset selected.
Faceless TikTok channels
Automate narration for faceless content by splitting long scripts into 4096-character chunks. The API handles the synthesis, allowing you to generate full videos without recording your own voice.
Use voices like Luna for soft storytelling or Aria for warm narration to keep viewers engaged without showing your face.
- Automate narration for faceless content
- Split long scripts into 4096-character chunks
- Use Luna for soft storytelling
Explainer videos
Generate clear, instructional audio using Sage’s calm tone. This voice is optimized for educational content, making complex topics easy to follow.
Combine with visual assets to create engaging explainers that retain viewer attention through consistent pacing and clarity.
- Use Sage for instructional audio
- Optimize for educational content
- Retain viewer attention
Product reviews
Use Felix’s energetic tone for promotional clips or Milo’s friendly style for conversational reviews. Both voices are designed to sound natural and engaging.
Generate reviews quickly by sending product descriptions to the API, ensuring consistent branding across all your review content.
- Use Felix for promotional clips
- Use Milo for conversational reviews
- Ensure consistent branding
Final selection tips
Selecting the right voice is the final step in building a reliable audio workflow for short-form video. The goal is consistency across your channel, ensuring that every clip feels like part of the same brand identity.
Avoid switching voices randomly. Pick one preset that aligns with your content style and stick with it to build audience recognition.
Match brand tone
Your voice should reflect the energy of your edits. Fast-paced cuts pair well with energetic presets like Felix, while calm storytelling suits Luna.
Keep the tone consistent with your visual style to avoid jarring transitions that break viewer immersion.
- Use Felix for high-energy promo clips
- Use Luna for soft, narrative-driven stories
- Use Atlas for neutral, factual updates
Consider target audience
Younger viewers on TikTok often respond better to bright, upbeat voices like Nova. Professional audiences may prefer the calm authority of Sage.
Align the voice personality with who you are trying to reach, not just what you are saying.
Test with real content
Use your $0.50 free trial to generate samples. Test each voice with your actual script length and pacing.
Compare side-by-side to hear how the voice interacts with your background music and sound effects before committing.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Matching voice personas to your target demographic helps align audio tone with viewer expectations.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- ISO 639 language codes
- Creative Commons Attribution 4.0
- LibriTTS-R paper (arXiv 2305.18802)
- TRON developer hub: TRC-20
- Tether (USDT)
- IETF RFC 7845: Ogg Opus
- W3C WAI: making audio and video accessible
- openai-python on GitHub
- LibriTTS-R corpus (OpenSLR 141)
- LibriVox public-domain audiobooks
- Node.js documentation
