Frequently asked questions
What is the best voice for TikTok?
Q: Which voice is best for TikTok? A: There is no single best voice; choose based on your content style. Use nova for upbeat reels, luna for storytelling, or atlas for neutral news. All eight voices support all 23 languages, allowing you to match the tone of your video precisely without manual recording.
How long can a TikTok voiceover be?
Each API request accepts up to 4096 characters. For longer scripts, split the text into parts, generate each segment, and join the audio files. This allows you to create voiceovers of any length suitable for TikTok, Reels, or Shorts.
Can I use the API for Reels?
Yes. The API returns standard audio files (mp3, wav, opus, flac, pcm) that work in any video editor. You can generate voiceovers programmatically and import them into Instagram Reels just like any other audio track without manual recording.
Does the API support Shorts?
Yes. The API generates high-quality synthetic speech compatible with YouTube Shorts. You can send text via POST request and receive the audio file to embed in your Shorts videos, supporting all 23 languages and eight voice presets.
How do I change the voice speed?
Include the speed parameter in your request body, with a value between 0.25 and 4.0. This adjusts the playback rate of the generated audio. The default speed is 1.0, but you can speed up or slow down the narration.
What format should I save the audio as?
Choose from mp3, wav, opus, flac, or pcm using the response_format field. MP3 is efficient for web use, while WAV or FLAC offers lossless quality. Select the format that best fits your video editing workflow and storage needs.
Can I generate voiceovers in bulk?
Yes, but rate limits apply. You can make 60 requests per minute per key. Process texts by splitting long scripts into 4096-character chunks. Requests are processed sequentially, so under load, a request may wait in a queue before generating.
Do I need coding skills?
Basic coding knowledge is helpful but not strictly required. You can use the OpenAI SDK or plain HTTP requests. The endpoint matches the OpenAI speech API shape, allowing you to use existing code snippets or tools like curl to generate audio.
How much does a TikTok voiceover cost?
Pricing is $10 per 1 million characters, or $0.01 per 1,000 characters. Failed requests are free. You can start with a $0.50 free trial for 7 days. Top-ups are via USDT (TRC-20) with no monthly plans or subscriptions.
Can I use the audio on YouTube?
Yes. The generated audio files are standard formats (mp3, wav, etc.) that you can upload to YouTube. Note that every file carries an inaudible watermark marking it as synthetic speech, which is standard for AI-generated content.
What if the text is too long?
Split your text into parts of up to 4096 characters. Generate audio for each part using separate API requests, then join the resulting audio files. This ensures high-quality synthesis and avoids character limit errors during generation.
How do I handle multiple languages?
The API supports 23 languages. Any voice can speak any supported language. If you do not set the language field, it guesses from the script (defaulting to English for Latin text). Explicitly set language for non-English Latin-script text.
Is the voice realistic?
Voices are based on licensed recordings from LibriTTS-R and LibriVox. They sound natural but carry an inaudible watermark indicating synthetic speech. They are not indistinguishable from human recordings, but they provide clear, consistent narration for your content.
Can I clone a voice for TikTok?
No. We do not offer custom voice cloning or uploading your own voice. You must choose from the eight available presets: aria, nova, luna, sage, felix, milo, atlas, or orion. Custom voice design is not part of this API.
Why use an API for TikTok
Generating voiceovers manually for every TikTok, Reel, or Short is slow and inconsistent. An API lets you produce audio programmatically, turning text into speech files on demand. This removes the bottleneck of recording, editing, and uploading voice tracks for each video.
You can feed your script directly into the API and receive an audio file in seconds. This method scales easily as your content volume grows, ensuring every video gets high-quality narration without manual effort.
Scale content creation
With a single API call, you generate audio for multiple videos. Process long scripts by splitting them into parts of up to 4096 characters and joining the results. This allows you to produce hours of content daily without hiring voice actors.
- Process up to 4096 characters per request
- Join audio files for longer scripts
- Generate audio in a few seconds for short phrases
Consistent brand voice
Select one of eight presets like Aria or Felix and use it for every video. The voice remains identical across all your content, building recognition. You do not need to re-record the same script to maintain consistency.
- Choose from 8 distinct voice presets
- Use the same voice for all videos
- Maintain tone across different languages
Automate workflows
Connect the API to your video editing pipeline. Pass text from your CMS or script generator directly to the speech endpoint. Receive the audio file and import it into your editor without manual intervention.
- Use standard HTTP requests
- Integrate with no-code tools like Zapier
- Automate text-to-audio conversion steps
Preparation steps
To start generating voiceovers for your TikTok, Reels, or Shorts content, you first need a ShortForm AI account. Create one via /get-api-key/ using Google or email. This gives you access to the dashboard where you manage your API resources.
Once registered, retrieve your unique API key from the dashboard. This key authenticates your requests to the API. You also need to select a voice preset that fits your content’s tone, such as 'nova' for upbeat shorts or 'atlas' for neutral news-style clips.
Sign up for account
Visit the /get-api-key/ page and sign up using Google or an email/password combination. Accounts are site-specific to play78j.com. Signing up with Google grants a $0.50 trial credit (approx. 50,000 characters) valid for 7 days. Email-only accounts start with zero credit.
- Google sign-up includes $0.50 free trial credit
- Email/password accounts start with no trial credit
- Accounts are independent per site
Get API key
After signing in, navigate to your dashboard to view your active API key. This key starts with tts- and authenticates your requests. You can issue a new key at any time; the old one stops immediately. Keep this key secure as it controls your usage and billing.
- Key format:
tts-... - Issue new key anytime; old one stops instantly
- One active key per account
Choose a voice
Select one of 8 presets based on your content’s tone. Options include female voices like 'aria' (warm narrator) or 'nova' (bright), and male voices like 'atlas' (neutral news) or 'orion' (deep documentary). Any voice can speak any of the 23 supported languages.
- 8 presets: aria, nova, luna, sage, felix, milo, atlas, orion
- All voices speak all 23 supported languages
- Voices are based on licensed LibriTTS-R recordings
Writing the script
Your script must fit within a single API call’s limit of 4096 characters. This constraint applies to the raw text input, including spaces and punctuation. If your voiceover exceeds this length, split it into multiple requests and join the resulting audio files afterward.
Clear punctuation ensures natural pauses in the generated speech. The TTS engine relies on commas and periods to determine rhythm. Avoid excessive ellipses or dashes, which may not render consistently across different voices or speeds.
Keep under 4096 characters
The API accepts up to 4096 characters per request. This limit is strict; exceeding it will cause the request to fail. For longer scripts, divide the text into chunks, generate each audio file, and concatenate them in your editing workflow.
Character counting includes spaces and punctuation. Each character costs $0.00001 (spaces and punctuation count as characters). Plan your script length carefully to avoid unnecessary costs from failed requests or excessive splitting.
- Limit input to 4096 characters per request
- Split long scripts into manageable chunks
- Concatenate audio files after generation
Use clear punctuation
Use standard commas, periods, and question marks to guide the TTS engine’s pacing. Avoid creative punctuation like em-dashes or multiple exclamation marks, as they may not produce the intended effect. Clear syntax leads to more predictable and natural-sounding voiceovers.
The engine does not support SSML. Test different punctuation styles with your chosen voice to find the best balance between speed and clarity.
- Use standard commas and periods
- Avoid em-dashes or multiple exclamation marks
- Test punctuation styles with your voice
Consider language tags
Always set the language parameter explicitly, especially for non-English text. If omitted, the API guesses the language from the script’s script (e.g., Latin defaults to English). Explicit tags prevent mispronunciation in mixed-language scripts.
Supported languages include English, Spanish, French, German, and 19 others. Ensure your script uses characters compatible with the selected language to avoid artifacts or errors in the generated audio.
- Set
languageparameter explicitly - Avoid reliance on automatic detection
- Check supported language list
Making the request
Generate audio by sending a POST request to the speech endpoint. Pass your API key in the Authorization header and define the model, voice, and input text in the JSON body. The API returns the audio file directly in the response body.
You can use the OpenAI SDK by pointing it to our base URL, or send a simple cURL command. This flexibility lets you integrate voice generation directly into your Python or JavaScript scripts without learning a new library.
Use cURL or SDK
Use the OpenAI SDK for TypeScript or Python by setting the base URL to https://api.play78j.com/v1. Alternatively, send a standard POST request with the model set to "tts-1" or "tts-1-hd". Include your text in the input field and your voice preset in the voice field.
- Base URL: https://api.play78j.com/v1
- Method: POST
- Auth: Bearer tts-...
Set speed parameters
Adjust the speed parameter between 0.25 and 4.0 to match your video’s pacing. Faster speeds suit quick cuts, while slower rates work for emotional storytelling. The API processes requests sequentially, so high-volume scripts may queue.
- Range: 0.25–4.0
- Default: 1.0
Handle response
The API returns the audio file directly in the response body. Save this file immediately, as it is not stored on our servers. If synthesis fails, you receive a refund, so you can retry without cost.
Processing the audio
Once the API returns the audio file, you can download it directly to your local machine. The response body contains the binary audio data in your chosen format, typically MP3 for compatibility with most editing software. This file is ready for immediate import into your timeline.
The generated audio is a standalone file. You will need to manually align the clip with your visual cuts in your editor. The file includes an inaudible watermark, but this does not affect playback quality or sync accuracy.
Download MP3 file
The API response returns the audio binary directly. Save this output as an .mp3 file. Most video editors accept this format without conversion.
If you used WAV or FLAC, the file will be larger but lossless. MP3 is recommended for most social media workflows to save storage space.
- Save the binary response as .mp3
- Verify file size before import
- MP3 is supported by all major editors
Sync with video
Import the audio file into your video editor. Align the waveform peaks with your visual transitions. Since the API does not provide frame-accurate timestamps, manual alignment is required.
Use the speed parameter (0.25–4.0) in your next request if the voiceover is too fast or slow for your visual pacing. Adjusting speed is more efficient than editing the audio clip.
- Align waveform with visual cuts
- Adjust speed parameter for pacing
- Manual sync is required
Add background music
Place your voiceover on an audio track above your background music. Ensure the music volume is low enough to not mask the speech. The synthetic nature of the voice is clear, so clean audio mixes work best.
Avoid complex sound effects that might clash with the synthetic voice. Every generated file carries an inaudible watermark that marks it as synthetic speech.
- Lower music volume under voiceover
- Keep mixes simple
- Watermark is safe for commercial use
Optimizing for short video
Short-form video relies on tight pacing. Adjusting the speed parameter (0.25–4.0) lets you compress voiceovers to fit 15–60 second clips without manual audio editing. Faster speeds increase energy, while slower rates improve clarity for instructional content.
Concise scripts maximize impact. With a 4096-character limit per request and roughly 25 characters generated per second, trim filler words. Clear punctuation ensures natural pauses, keeping the audio tight for Reels and Shorts.
Adjust speed for pacing
Use the speed parameter to control playback rate. Values between 1.0 and 2.0 suit fast-paced TikTok trends, while 0.8–1.0 works for storytelling. Requests are processed sequentially, so longer audio may queue under load.
- Set speed: 1.5–2.0 for rapid hooks
- Set speed: 0.8–1.0 for clear instructions
- Monitor queue times during peak usage
Use concise text
Every character counts toward your $0.01 per 1,000 characters cost. Remove redundant phrases to keep scripts under 4096 characters. Clear punctuation helps the engine generate natural pauses, reducing the need for post-production cuts.
- Limit input to 4096 characters
- Remove filler words
- Use commas for natural pauses
Test different voices
Try different presets like nova for upbeat energy or atlas for neutral news-style delivery. Each voice carries an inaudible synthetic watermark. Test multiple options to find the tone that matches your brand identity.
nova: bright, upbeatatlas: neutral, news-stylesage: calm, instructional
Common mistakes
Developing TikTok voiceovers programmatically requires precision to avoid wasted credits and broken pipelines. Many creators encounter avoidable errors that halt automation or produce unusable audio files. Understanding these common pitfalls ensures your API integrations run smoothly and cost-effectively.
By addressing character limits, language tags, and error handling early, you can maintain consistent audio quality across hundreds of short-form videos without manual intervention or unexpected billing surprises.
Ignoring character limits
The API enforces a strict 4096-character limit per request. Exceeding this causes the request to fail. If your script exceeds this length, you must split it into smaller parts and join the resulting audio files manually or via your automation tool.
- Split long texts into chunks of up to 4096 characters.
- Join the resulting audio files to create a seamless voiceover.
Wrong language code
If the language parameter is omitted, the system guesses the language based on the script. For non-English Latin-script text, this often defaults to English, leading to unnatural pronunciation. Always set the language field explicitly to ensure accurate speech synthesis for your target audience.
Not handling errors
Failed requests are free, but ignoring error responses wastes time. Ensure your script checks the HTTP status code and response body. If synthesis fails, retry or log the issue without charging your balance. Proper error handling prevents silent failures in your video generation pipeline.

Tools integration
Integrate the API into your existing creation pipeline to automate voiceover generation. Since the endpoint accepts standard JSON, you can plug it directly into your preferred tools without needing a custom plugin, mobile SDK, or our own SDK.
This approach keeps your workflow headless and programmatic, allowing you to trigger audio synthesis automatically when a new script is ready.
Connect to video editor
Most modern editors accept external audio files. Generate the MP3 via API first, then import it as a track. This decouples voice synthesis from video rendering, allowing you to swap voices without re-rendering the entire timeline.
- Generate audio first, then import to editor
- Swap voices without re-rendering video
- Supports MP3, WAV, OPUS, FLAC, PCM
Use Zapier or Make
Use Zapier or Make.com to trigger API calls. When a new row is added to a spreadsheet or a form is submitted, send the text to the TTS endpoint. The resulting audio URL or file can be saved to cloud storage for later use in your editor.
Set up error handling to retry failed requests, ensuring no script is lost due to temporary API issues.
- Trigger on form submission or spreadsheet update
- Auto-save generated audio to cloud storage
- Handle retries for failed synthesis
Script automation
Write a simple Python or Node.js script to fetch scripts from your CMS, split them if over 4096 characters, and send them to the API. Loop through the responses to join audio files sequentially.
This ensures consistent pacing and allows you to batch-process multiple videos at once, scaling your content output significantly.
- Split long texts into 4096-char chunks
- Join audio files sequentially
- Batch process multiple videos
Cost management
Cost efficiency is critical for high-volume content creation. You are charged $0.01 per 1,000 characters, so monitoring your balance and usage helps prevent unexpected stops. Failed requests are free, but you must track them to ensure accurate billing.
Top-ups use USDT on the TRON network (TRC-20) only. Minimum $5 transfers are credited automatically within minutes. There are no monthly plans or subscriptions, so your balance does not expire.
Monitor character usage
Track your remaining characters in the dashboard. Since spaces and punctuation count, concise scripts reduce costs. Use the API usage stats to predict monthly spend based on your posting schedule.
- Check balance in /get-api-key/
- Characters are charged at $0.01/1k
- Failed requests are free
Top up USDT
Send USDT via TRC-20 to your wallet address. Transfers are credited automatically within minutes. No invoices, no VAT, and no minimum spend beyond $5.
- Use TRON network (TRC-20)
- Minimum top-up is $5
- Credit arrives in minutes
Track failed requests
Synthesis failures result in a full refund. Monitor your request history to identify patterns. Ensure your API key is active, as only one key is valid at a time.
- Refunds are automatic
- Check request status code
- No hidden fees
Voice selection
Selecting the right voice is critical for TikTok engagement. Our eight presets cover distinct tones, from the warm clarity of Aria to the deep documentary style of Orion. Each voice has a distinct tone, ensuring your voiceover matches the video’s mood without manual direction.
The API returns synthetic speech with an inaudible watermark, confirming the audio is AI-generated. This is standard for platform compliance. Choose voices that align with your niche, whether it’s the energetic promo style of Felix or the calm instructional tone of Sage.
Choose based on tone
Match the voice to your content type. Use Aria for clear narrations or Luna for storytelling. For fast-paced shorts, Felix’s energetic tone works best. Milo offers a friendly, conversational vibe suitable for tutorials.
Each voice has distinct characteristics derived from licensed recordings. Select based on whether you need neutrality (Atlas) or depth (Orion). There is no emotion tagging, so the tone is fixed per voice.
- Aria: Warm, clear narrator
- Felix: Energetic, promo
- Milo: Friendly, conversational
- Atlas: Neutral, news
Test multiple voices
Generate short samples for each voice using the API. Compare how each interprets your script. Voices are consistent across languages, so test with your actual script text to check natural flow.
Use the 8 presets to find the best fit. You can switch voices instantly without retraining. Generate a few seconds of audio for each candidate to evaluate clarity and pacing.
- Generate 5-second samples
- Compare clarity and tone
- Switch voices instantly
Match brand identity
Consistency builds recognition. Pick one primary voice for your series. Use secondary voices for different segments if needed. Ensure the tone aligns with your visual style and target audience.
Avoid mixing too many voices in one video. Stick to one preset for brand coherence. Use the character limit to keep scripts concise, ensuring the voice delivers the message clearly.
- Pick one primary voice
- Maintain tone consistency
- Limit voice changes per video
Language support
The API guesses the language from the script by default, assuming English for Latin text. However, relying on auto-detection can lead to incorrect accents if the input text contains ambiguous phrasing. For consistent results, always specify the language parameter in your request.
Explicitly setting the language ensures the voice model applies the correct pronunciation rules and intonation patterns. This is critical for non-English content, preventing the AI from defaulting to an English accent on words that look similar across languages.
Set language explicitly
Include the language field in your JSON body with the ISO code (e.g., "language": "fr"). This overrides the automatic detection and guarantees the voice speaks the intended language correctly.
This is especially important for texts that might mix scripts or contain loanwords. Without an explicit tag, the engine may misinterpret the language context.
- Always define the
languagefield for non-English content. - Use standard ISO codes like
en,de,ja,es. - Prevents accidental English pronunciation on Latin-script foreign text.
Handle multi-language
Each API call processes a single text input. To create a video with multiple languages, split your script into separate requests for each language segment.
Concatenate the resulting audio files in your video editor. The engine does not handle language switching within a single request.
Ensure your text segments are clearly separated to avoid pronunciation artifacts at the boundaries.
- Split multi-language scripts into separate API calls.
- Join audio files in your editing software.
- Maintain consistent voice across segments for continuity.
Check supported list
The API supports 23 languages, including English, Spanish, French, German, Chinese, Japanese, and Arabic. If your target language is not in this list, synthesis may produce unintelligible results.
Unsupported languages include Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, and Bengali. Verify your needs before integrating.
If language is omitted, Latin script defaults to English. Use the GET /v1/voices endpoint to confirm available options.
- Supported: Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish.
- Not supported: Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali.
- Default language for Latin text is English if not specified.
Quality control
Quality control ensures your TikTok voiceovers sound professional and ready for viral distribution. By listening to the generated audio and verifying technical details, you maintain consistent brand identity across all your short-form content.
Listen to output
Play the generated MP3 file to check natural flow and pacing. Adjust speed parameters if the delivery feels too fast or slow for your video's rhythm. Ensure the voice matches the intended tone of your content.
Check for artifacts
Listen for robotic pauses or unnatural intonations, especially around complex punctuation. Since requests are processed sequentially, long texts may show slight timing variations between split parts. Trim or re-generate segments that disrupt the flow.
Verify watermark
Every file contains an inaudible watermark marking it as synthetic speech. This is expected and does not affect audio quality. You do not need to remove it, but be aware it is present in the final file.
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.
