lovekit
The python client for lovekit, the Arkitekt service for live audio and video. An app (an agent: one client, user and organization) declares the live sources it offers, its streams, and lovekit keeps them, hands out the tokens to publish and watch them, and says which are live. The media itself flows over WebRTC through a LiveKit media server.
Installation
pip install lovekit # or: pip install "arkitekt[rekuest,lovekit]"
Usage
Every lovekit operation is a method of the Lovekit client, in a blocking and
an a-prefixed async flavour: declare_streams, publish_token, watch_stream,
watch_agent, get_stream, is_stream_live, list_streams, get_agent, ….
Publishing streams
lovekit.publish(...) declares an app's streams and keeps them published:
from lovekit import VideoStream, to_rgba
async with lovekit.publish([VideoStream("camera", title="Main camera")]) as live:
print(live.streams) # the declared streams (@lovekit/stream)
while True:
await live["camera"].send(to_rgba(frame)) # (height, width, 4) uint8
- The declaration is the app's whole set. A stream keeps its id over restarts (it is found by its key), one the app no longer names is removed, and a restart that declares the same changes nothing. So a client publishes once at a time.
- Every stream is a track of the agent's LiveKit room, named by its key. A track is made on the first frame and again when the frame size changes; the last frame is sent again while no new one comes, so a late viewer sees a picture.
- It stays live. When the room drops it is joined again (fresh token, backoff) and
every track published again. When another process of the same agent connects, or
the participant is removed, this one stops for good (
live.stopped,await live.wait_stopped()) instead of taking the room back. - Audio is an
AudioStream("microphone"), sent int16 samples.
In an arkitekt app
The client is injected by annotation. Hold the streams in a state, so the UI
shows them (in a blok: <foreach items="@self.Live.streams" let="#stream">
<StructureDisplay value="@stream" /></foreach>), and publish from a background
hook:
from dataclasses import dataclass, field
from arkitekt import App, run
from lovekit import VideoStream, lovekit_service, to_rgba
from lovekit.api.schema import Stream
from lovekit.lovekit import Lovekit
app = App("camera", "0.1.0", services=[lovekit_service])
@app.state
@dataclass
class Live:
streams: list[Stream] = field(default_factory=list)
@app.startup
def no_streams_yet() -> Live:
"""Every state starts from a startup hook: empty, until the streams are declared."""
return Live()
@app.background
async def stream_camera(lovekit: Lovekit, live: Live) -> None:
async with lovekit.publish([VideoStream("camera")]) as published:
live.streams = list(published.streams)
while not published.stopped.is_set():
await published["camera"].send(to_rgba(grab_frame()))
if __name__ == "__main__":
run(app)
Streams and agents travel between actions by id (@lovekit/stream,
@lovekit/agent), so an action can take and return them directly.
Watching
watch_stream(stream) and watch_agent(agent) return a token that may subscribe in
the agent's room, but not publish; the stream is the track named by its key.
is_stream_live(stream) says whether it is published right now.
Development
The generated API (lovekit/api/schema.py) comes from the documents in
graphql/; see graphql.config.yaml.
uv run pytest -m "not integration" # no server needed
uv run pytest -m integration # a real lovekit + livekit via dokker
See RELEASING.md for how versions are cut.
1 maintainer
Metadata
1 maintainer
Release files for lovekit 4.0.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| lovekit-4.0.0.tar.gz | 20.4 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| lovekit-4.0.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 42.5 kB
Release files / lovekit-4.0.0.tar.gz
| Download URL | lovekit-4.0.0.tar.gz |
|---|---|
| Size | 20.4 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
09a04eb867fdc2efc15bd866972232150726b5d11f9072319bc45e162dde6f2b
|
|
BLAKE2b-256 checksum How to use checksums |
d583c080382aa1754dbf4d8e0a3f95cf9607fca1d95cfef676a58154e00e0558
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
uv/0.13.1 {"installer":{"name":"uv","version":"0.13.1","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
|
Release files / lovekit-4.0.0-py3-none-any.whl
| Download URL | lovekit-4.0.0-py3-none-any.whl |
|---|---|
| Size | 22.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
02332ac2081d434d9cf11c77440fa4da69303606553d56c3b806ce1321adab7b
|
|
BLAKE2b-256 checksum How to use checksums |
8dc663c680b0fb7daefff97014ec3830d4ebd2d328c9b642be89e8b40f952d07
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
uv/0.13.1 {"installer":{"name":"uv","version":"0.13.1","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
|
Security for lovekit 4.0.0
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.
