Skip to main content

Lip Service

Talk is cheap. Singing is free.

Hear it in your browser · Write a song · GitHub

Sixteen voices that sing what you write and say what you type. It is an instrument as much as a speech engine, and it runs on your own CPU: no GPU, no service, no dependencies.

pip install lipservice
lipservice say "Hello there."

The first call fetches the models, about 56 MB, and keeps them. The sound is made on your machine; each run checks lipservice.dev's server for changed models first.

From Python

import lipservice as lip

lip.say("Hello there.").save("hello.wav")
lip.say("And now, a little faster.", voice="ben", rate=1.15).play()

song = """
bpm 90
Twinkle C4:1 C4:1
twinkle G4:1 G4:1
little A4:1 A4:1
star G4:2
"""
lip.sing(song, voice="cora").save("star.wav")

say and sing return a Clip: save(path) writes a WAV file, play() plays it, samples is the sound as floats at 24 kHz, and a notebook shows it as a player.

Voices

lipservice voices lists them: ada, ben, cora, dev, eve, finn, gwen, hal, ivy, jack, kate, leo, mae, ned, opal, pete.

Two qualities

quality="hifi" is the default. quality="fast" is quicker. Both are in the one package.

A song is text

Save this as song.txt:

bpm 90
amazing D4:1 G4:2 B4:0.5+G4:0.5
grace B4:2
rest 1
how A4:1
sweet G4:2
the E4:1
sound D4:2

A word a line, and a note for each of its syllables: NOTE:BEATS, with the note as C4, F#3, Bb2 or a MIDI number. + joins notes on one syllable. rest BEATS is silence and bpm N sets the tempo. Songs are sung where they are written: shift=-12 puts one written for a woman in a man's range. The whole format, and what the engine refuses, is in SCORE.md in the repository.

From the command line

lipservice say "Hello there." -o hello.wav --voice ben
lipservice sing song.txt --voice cora --motion vibrato
lipservice voices
lipservice fetch

Without -o the sound is played.

Choosing where it plays

lipservice devices lists this machine's sound outputs: an id, a name, and * on the system's default. --device ID plays through one of them, and so does clip.play(device=ID), with lip.devices() for the list. Two outputs can share a name; the id tells them apart.

For a model: the MCP server

lipservice mcp serves speech and song to a language model over the Model Context Protocol, on stdin and stdout. It offers four tools: voices, devices, say and sing. A client that takes MCP servers as JSON is given:

{
  "mcpServers": {
    "lipservice": {"command": "lipservice", "args": ["mcp", "--out", "/path/to/a/folder"]}
  }
}

Each say and sing puts its sound where the call says: device, the id of an output, plays it there and returns when it has played; out writes a WAV file. A call that gives neither writes a new file in the server's folder, which is --out DIR or the LIPSERVICE_OUT variable. With no folder, such a call is refused. Nothing is played that a call did not name a device for.

One clip plays at a time. A call the client cancels stops playing.

The repository's plugin/ folder is this server and a skill that teaches the score format, packaged for Claude Code, Codex, Gemini CLI and Antigravity. For Claude Code the repository is a plugin marketplace: claude plugin marketplace add it, then claude plugin install lipservice@lipservice; for Codex, codex plugin marketplace add it. Each needs lipservice installed where the client runs.

License

AGPL-3.0-only. A commercial license is available for use the AGPL does not suit: see https://lipservice.dev.

Metadata

Release files for lipservice 0.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for lipservice 0.1.0
File Size Uploaded Provenance Actions
lipservice-0.1.0.tar.gz 261.1 kB Without provenance Details for lipservice-0.1.0.tar.gz

Built distributions (wheels)

Table of built distributions (wheels) for lipservice 0.1.0
File Actions
lipservice-0.1.0-py3-none-musllinux_1_2_x86_64.whl Python 3 none Linux musl 1.2+ x86-64 Details for lipservice-0.1.0-py3-none-musllinux_1_2_x86_64.whl
lipservice-0.1.0-py3-none-musllinux_1_2_aarch64.whl Python 3 none Linux musl 1.2+ ARM64 Details for lipservice-0.1.0-py3-none-musllinux_1_2_aarch64.whl
lipservice-0.1.0-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl Python 3 none Linux glibc 2.17+ x86-64 Details for lipservice-0.1.0-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl
lipservice-0.1.0-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl Python 3 none Linux glibc 2.17+ ARM64 Details for lipservice-0.1.0-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl
lipservice-0.1.0-py3-none-macosx_11_0_arm64.whl Python 3 none macOS 11.0+ ARM64 Details for lipservice-0.1.0-py3-none-macosx_11_0_arm64.whl
lipservice-0.1.0-py3-none-macosx_10_12_x86_64.whl Python 3 none macOS 10.12+ x86-64 Details for lipservice-0.1.0-py3-none-macosx_10_12_x86_64.whl

Total release size: 4.7 MB

Release files / lipservice-0.1.0.tar.gz

Download URL lipservice-0.1.0.tar.gz
Size 261.1 kB
Tags Source
SHA-256 checksum
How to use checksums
8fbcfc22c709bf2d6d5c3c1eb9174dc74d5e2d5c2365581a15924c0c62d24464
BLAKE2b-256 checksum
How to use checksums
f07a04e9f327469bb926e0c25485aabef5db4eeb250a477528c9ac68bebeed44
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.9

Release files / lipservice-0.1.0-py3-none-musllinux_1_2_x86_64.whl

Download URL lipservice-0.1.0-py3-none-musllinux_1_2_x86_64.whl
Size 834.6 kB
Tags Linux musl 1.2+ x86-64 Python 3
SHA-256 checksum
How to use checksums
860d914c1c5015bca2141c938b59f338a4babb5e4349b52a0814042ec4a504b1
BLAKE2b-256 checksum
How to use checksums
eea1cd540705a32dcb4dcde287b812f875df1fe864dd556f4b7ee62e78df765f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.9

Release files / lipservice-0.1.0-py3-none-musllinux_1_2_aarch64.whl

Download URL lipservice-0.1.0-py3-none-musllinux_1_2_aarch64.whl
Size 783.2 kB
Tags Linux musl 1.2+ ARM64 Python 3
SHA-256 checksum
How to use checksums
81b73323fd84b2ef11bbcd457b8b9e3004a76fcb0f7bf058b3a031fe3f1cf440
BLAKE2b-256 checksum
How to use checksums
906bd872c3e072fca20a0feab633c393a5b5a262caca132d56f463ae2abdcc26
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.9

Release files / lipservice-0.1.0-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl

Download URL lipservice-0.1.0-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl
Size 748.3 kB
Tags Linux glibc 2.17+ x86-64 Python 3
SHA-256 checksum
How to use checksums
3582122a0f5ebc4e814c120f15835a3804116cbc15d0ba0cf719c3f96812a913
BLAKE2b-256 checksum
How to use checksums
91c1735c60ef42df36cfc83ca5ba3a15060f130dc7f52e77bd77e83f5aacb5c3
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.9

Release files / lipservice-0.1.0-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl

Download URL lipservice-0.1.0-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl
Size 716.6 kB
Tags Linux glibc 2.17+ ARM64 Python 3
SHA-256 checksum
How to use checksums
29e1e773530bf92265d789842d846647febfd335d4776acae918eb3b4384f699
BLAKE2b-256 checksum
How to use checksums
1fee873f0a18c2453aba7b85ebf5939c39f85e41f9bdafdb51537bb4c927aac6
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.9

Release files / lipservice-0.1.0-py3-none-macosx_11_0_arm64.whl

Download URL lipservice-0.1.0-py3-none-macosx_11_0_arm64.whl
Size 666.2 kB
Tags Python 3 macOS 11.0+ ARM64
SHA-256 checksum
How to use checksums
92fa9584fc3d18a0ff4032a506fa3f2de3c6fb74837e566162194fd717145d52
BLAKE2b-256 checksum
How to use checksums
12902d66a9b810d2a44d5b1a5f33eed6f4ac61dc65945a2866e903cf5fccb387
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.9

Release files / lipservice-0.1.0-py3-none-macosx_10_12_x86_64.whl

Download URL lipservice-0.1.0-py3-none-macosx_10_12_x86_64.whl
Size 719.1 kB
Tags Python 3 macOS 10.12+ x86-64
SHA-256 checksum
How to use checksums
de59e4d3d3c0b8a17f8543da747214b7e181e357051f9685fbcfccde87e03153
BLAKE2b-256 checksum
How to use checksums
ff37dd297eed601d23b7b5cdc98fda3df6fb5633da6ae3ffbe30ec04f069bef7
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.9

Security for lipservice 0.1.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.1.0 This release

7 release files No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page