Xiaomi MiMo Authorized reference voice synthesis
MiMo V2.5 TTS Voice Clone
Draft approved pickup lines from a recording you have permission to use for voice cloning.
Xiaomi describes MiMo V2.5 TTS Voice Clone as speech synthesis guided by a reference recording. Start with a clean sample and a short approved script, then let the speaker or authorized reviewer judge the result.
- Authorized reference voice synthesis
- Speech audio output
- Non-streaming WAV in Playground
The Playground requires a reference recording and the words to synthesize. Use a recording of your own voice, or obtain the speaker's explicit permission covering voice cloning and the intended use. TokenHot's API guide accepts MP3 or WAV data URLs and limits the Base64-encoded string to 10 MB.
Try this model
Generate speech from text using this model.
mimo-v2.5-tts-voicecloneInput
Generated result
Enter a prompt, choose your settings, and generate.
USD /1M tokens
Pricing
| Group | Input | Output |
|---|---|---|
| default | $0 | $0 |
Public list prices are shown in USD. Final charges may vary by account group and usage tier.
Overview
What is MiMo V2.5 TTS Voice Clone?
The Playground requests a complete WAV. The TokenHot guide puts the sample in audio.voice and the script in the assistant message. The API can accept style instructions in an optional user message.
Start with a permitted recording
Use a clear sample from a speaker who approved cloning for this project.
Compare pronunciation and delivery
Listen to the generated line against the script and the permitted reference.
Keep the approval scope with the asset
Track the approved purpose and reviewer alongside the clip.
From brief to deliverable
Put MiMo V2.5 TTS Voice Clone to work
Voice production teams
Draft two approved pickup lines
Use the same reference for each separate line, then compare how each take fits the script.
- Tools you need
- The TokenHot Playground and a listener who can review the script and recording.
- Keep in mind
- Get approval for the generated lines before use. Keep the authorization tied to the specific project.
View a task brief
Attach the authorized reference recording and paste the approved pickup line into Script. Compare pacing with the surrounding audio.
Audio editors
Replace one line in an approved narration
Generate a candidate for the changed sentence and place it in a copy of the edit timeline for review.
- Tools you need
- The TokenHot Playground and a listener who can review the script and recording.
- Keep in mind
- Compare levels and pacing at the edit point. Keep the original narration available for rollback.
View a task brief
Attach the authorized reference recording and paste the exact replacement sentence into Script. Check how it joins the existing narration.
Training and enablement teams
Review a voice update for internal training
Generate one short section and review it before updating the remaining lesson audio.
- Tools you need
- The TokenHot Playground and a listener who can review the script and recording.
- Keep in mind
- Use the recording only within its agreed scope. Verify technical terms and numbers with the lesson owner.
View a task brief
Attach the authorized recording and paste the approved training paragraph into Script. Listen for mispronunciations and pauses before accepting it.
Capabilities in context
Key features
| Feature | What you get | Why it matters |
|---|---|---|
| Exact model ID | mimo-v2.5-tts-voiceclone | Use this ID in the TokenHot request. |
| API endpoint | POST /v1/chat/completions | Follow the speech request fields in the model-specific guide. |
| Playground reference | One required MP3 or WAV sample up to 7 MiB | Choose an authorized recording before generating. |
| API reference field | audio.voice data URL | TokenHot limits the Base64-encoded string to 10 MB. |
| Playground output | Non-streaming WAV | TokenHot's API guide describes streaming compatibility mode as returning after inference completes. |
Xiaomi describes the three MiMo speech models. TokenHot's guide and Playground cover the route and its interface. The scene illustrations were made with ImageGen and do not represent audio results. TokenHot model API · Xiaomi MiMo V2.5 TTS release
Audience & fit
Where MiMo V2.5 TTS Voice Clone fits
Voice production teams
Two candidate lines for the authorized speaker or reviewer to approve.
Audio editors
A replacement take for comparison with the surrounding narration.
Training and enablement teams
A candidate update with script, pronunciation, and rights checks.
Cost context
Plan the cost of a useful result
Use the current route price
Check the live TokenHot price table and the rates that apply to your account.
Listen to a short pilot first
Review a representative passage before generating the full script or a batch of lines.
Count revisions and review
Include regenerated lines and the listening time needed to approve usable audio.
Use the current pricing table for available rates and account groups. Model IDs and prices come from TokenHot’s catalog; provider pricing and subscriptions are separate.
Blog
Related Articles
FAQ
Frequently asked questions
What is MiMo V2.5 TTS Voice Clone best suited for?
MiMo V2.5 TTS Voice Clone is suited to the input, output, and capability types listed on this page. Test production workloads before choosing it for a critical system.
How is MiMo V2.5 TTS Voice Clone priced?
Pricing is shown in the pricing section above. Actual cost depends on usage volume, account group, and the request payload.
How can I call MiMo V2.5 TTS Voice Clone?
Use the model ID shown above with one of the supported API protocols. Code examples are provided when usage data is available.
What permission do I need for the recording?
Use your own voice or get explicit permission for voice cloning and the intended use. Have the speaker or authorized reviewer approve the resulting lines.
What format does the API accept for the sample?
Use an MP3 or WAV data URL in audio.voice. TokenHot's guide caps the Base64-encoded string at 10 MB.
Can I choose a preset voice instead?
This route takes a reference recording. For a preset voice, use mimo-v2.5-tts.
Does streaming return low-latency partial audio?
TokenHot documents compatibility behavior for this route, with the result returned after inference completes. The Playground requests WAV without streaming.
Can the same voice be guaranteed across all lines?
Listen to each candidate and compare it with the authorized reference and adjacent lines before release.
Get started
Build with MiMo V2.5 TTS Voice Clone
API