Self-hosting
Run Kokoro TTS locally for free voiceovers
Kokoro is an 82-million-parameter open-weight text-to-speech model under the Apache 2.0 license. It runs on your own machine with no account, credits or per-character fees, which makes it a real option for creators who narrate a lot.
- Platform
- Kokoro TTS
- Best for
- Developers and creators who want free, self-hosted TTS without per-character fees, and who can handle model setup.
- Hands-on
- Controlled benchmark pending
Quick answer
Install the kokoro Python package and espeak-ng, create a pipeline with a language code, pick a voice and generate WAV files. It is free for commercial use under Apache 2.0. Choose ElevenLabs instead when you need voice cloning, a large voice library or the most expressive delivery.
Starting point: Free and open source (Apache 2.0 license). · Free tier: Yes — Entire model is free (Apache 2.0).
This evaluation is built on verified documentation and official pricing data, paired with a repeatable test workflow you can run directly in your own project.
Worth evaluating when
- Creators who publish a lot of narration and want to avoid per-character fees
- Developers adding text-to-speech to scripts, apps or automations
- Anyone who wants voiceovers generated on their own hardware
Check before paying
- It is a model, not an app: setup needs Python and a terminal
- The official voice set is smaller than commercial libraries, and emotional range is below top commercial engines
- No built-in voice cloning; that is where paid tools like ElevenLabs still lead
What Kokoro is
Kokoro-82M is an open-weight text-to-speech model published by hexgrad. With 82 million parameters it is small enough to run on ordinary hardware, and its weights are released under the Apache 2.0 license, so you can use it in personal and commercial projects without paying per character.
Because it runs locally, there is no account, no monthly credit allowance and no upload of your scripts to a third-party service.
Install and generate your first file
The project's quick start uses Python: install the package with pip install kokoro>=0.9.4 soundfile, then install espeak-ng (for example apt-get install espeak-ng on Debian or Ubuntu), which Kokoro uses as a fallback for English and for some other languages.
In Python, import KPipeline from kokoro and create a pipeline with a language code, for example KPipeline(lang_code='a') for American English. Pass your text and a voice such as af_heart to the pipeline, then write each generated audio segment to a WAV file with soundfile.
Language codes in the documentation include a (American English), b (British English), e (Spanish), f (French), h (Hindi), i (Italian), j (Japanese), p (Brazilian Portuguese) and z (Mandarin Chinese). Some languages need extra packages; check the repository README.
A practical voiceover workflow
Split long scripts into paragraphs so you can regenerate a single line without re-rendering the whole video. Keep a text file of pronunciations for brand names and technical terms, and adjust the speed parameter if narration feels rushed.
Import the WAV files into your editor (CapCut, DaVinci Resolve or Descript) and add music and captions there.
Running Kokoro on a VPS
If you want narration generated automatically, for example by an n8n workflow or a script that turns articles into audio, run Kokoro on an always-on Linux server. A small model like Kokoro can run on CPU, so you do not need a GPU server to start; measure generation time with your own scripts before sizing up.
Keep any HTTP wrapper you add private or behind authentication, and pin the package version you tested.
When ElevenLabs is worth paying for
Kokoro covers clean narration for free. Paid engines such as ElevenLabs add voice cloning, a much larger voice library, dubbing and more expressive delivery. ElevenLabs has a free plan with 10,000 credits a month and paid plans from $6/month; check voice and commercial rights for your channel. The Kokoro vs ElevenLabs comparison covers the trade-off.
Run it 24/7 on your own server
Kokoro TTS runs on your own machine, so it stops answering when your laptop sleeps. A small Linux VPS with Docker keeps it online for scripts, agents and teammates, with a fixed monthly cost instead of per-token markup.
- Start with the smallest plan that runs Docker; upgrade only if you add users or local models.
- Keep the gateway bound to localhost or behind a firewall, and store provider keys as environment variables.
- Monthly billing lets you test before committing to a long term.
Some links are affiliate links. We may earn a commission if you buy through them. Payment never changes our rankings or editorial verdicts.
Related comparisons
Compare Kokoro TTS against the tools creators most often weigh it against.
Related guides
More Voice workflows from the same verified catalog.
Run a same-input test
Install
Install kokoro and soundfile with pip and espeak-ng with your package manager.
Generate a sample
Create KPipeline(lang_code='a'), generate one paragraph with the af_heart voice and save it as WAV.
Test your vocabulary
Run a script full of brand names and numbers and note which words need respelling.
Time a full script
Generate a complete 5-minute script and record how long it takes on your hardware.
Compare
Generate the same paragraph with a paid engine's free plan and decide whether the difference matters for your channel.
Official sources
These external sources support the product and pricing record. Recheck them before purchasing because plans and limits change.
Frequently asked questions
Is Kokoro TTS free for commercial use?
Kokoro's weights are released under the Apache 2.0 license, which allows commercial use. Hosted services built on Kokoro may charge for their own service.
Does Kokoro need a GPU?
It is a small 82M-parameter model and can run on CPU; a GPU speeds up long scripts. Test on your own hardware before buying a server.
Can Kokoro clone my voice?
The official model ships with a set of built-in voices and does not include voice cloning. Use a paid tool such as ElevenLabs if you need a cloned voice.
Which languages does Kokoro support?
The documentation lists language codes for American and British English, Spanish, French, Hindi, Italian, Japanese, Brazilian Portuguese and Mandarin Chinese; some need extra packages.