kymoscope

Stream live, end-to-end encrypted telemetry to Kymoscope from Python.

Requires Python 3.9 or later.

pip install kymoscope

Example

import math, time
from kymoscope import LiveStream

API_KEY = "kymoscope_..."   # Account -> API & real-time streaming, or a project's API key
start = time.time()

# state_path keeps this sender's identity, so a restart needs no new approval
with LiveStream(API_KEY, "Rig A", state_path="kymoscope-state.json") as stream:
    print("stream id:", stream.stream_id)
    while True:
        t = time.time() - start
        stream.send([{"t": t, "signals": [
            {"name": "speed", "unit": "m/s", "value": 50 + 10 * math.sin(t)},
        ]}])
        time.sleep(1 / 60)

In an end-to-end encrypted project, the first run waits until a project member's browser approves the sender; after that it publishes on its own.

Keep every sample when the connection drops

with LiveStream(API_KEY, "Rig A", state_path="kymoscope-state.json",
                buffer_path="kymoscope-buffer", buffer_max_mb=2048) as stream:
    ...

With buffer_path (a folder), the stream always sends, also with nobody watching, so a cloud recording keeps everything. While the relay is unreachable it keeps the encrypted messages on disk, then sends them in order when the relay is back, with the bandwidth live data leaves spare. Viewers see that data in the recording, not in the live view. A crash can send a few messages twice. buffer_max_mb caps the disk used; when it is full, new samples are dropped and counted in stream.buffer_dropped. block_format=True sends the compact format (4–8× less data); buffer_path turns it on.

The headless viewer (kymoscope.viewer) needs msgpack: pip install 'kymoscope[viewer]'.

Full documentation: https://kymoscope.com/docs/live