mirror of
https://github.com/encounter/adk-python.git
synced 2026-07-09 18:19:28 -07:00
chore: bidi streaming example on run live fast api
This sample connects to adk agent api server with websocket, to exchange streaming audio/text. PiperOrigin-RevId: 805531120
This commit is contained in:
committed by
Copybara-Service
parent
707c1a7bd4
commit
3c433b7168
Binary file not shown.
Binary file not shown.
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,26 @@
|
||||
# What's this?
|
||||
|
||||
This is a sample that shows how to start the ADK api server, and how to connect
|
||||
your agents in a live(bidi-stremaing) way. It works text and audio input, and
|
||||
the response is always audio.
|
||||
|
||||
## Prerequisite
|
||||
|
||||
- Make sure you go through https://google.github.io/adk-docs/streaming/
|
||||
|
||||
## Instruction for this sample
|
||||
|
||||
- The audio libraries we used here doesn't have noise cancellation. So the noise
|
||||
may feed back to the model. You can use headset to avoid this or tune down
|
||||
voice volume, or implement your own noise cancellation logic.
|
||||
- Please ensure you grant the right mic/sound device permission to the terminal
|
||||
that runs the script. Sometimes, terminal inside VSCode etc dones't really work
|
||||
well. So try native terminals if you have permission issue.
|
||||
- start api server first for your agent folder. For example, my anents are
|
||||
locoated in contributing/samples. So I will run
|
||||
`adk api_server contributing/samples/`. Keep this running.
|
||||
- then in a separate window, run `python3 live_agent_example.py`
|
||||
|
||||
## Misc
|
||||
|
||||
- Provide a few pre-recorded audio files for testing.
|
||||
Reference in New Issue
Block a user