Capture from Pipecat

setup drops a pipeline recorder: caller and agent on separate channels, to score offline.

The recording you capture feeds a five-dimension conversation-test: separated caller and agent channels are what the Conversation (turn-taking) and Speech (timing) dimensions read, and the same file drops into hotato test run beside a transcript and trace for the Outcome and Policy checks.

Stack	Automatic pull	Separate channels	Contract-grade default	Network use
Vapi	Yes: `capture` fetches by call id	Yes, out of the box	Yes	One call: fetch the recording
Twilio	Yes: `capture` fetches by recording SID	No by default: must request `RecordingChannels=dual` at creation	No, opt-in	One call: fetch the recording
Retell	Yes: `capture` fetches by call id	Yes, once call recording is turned on	Opt-in: enable recording first	One call: fetch the recording
LiveKit	No: egress runs in your own infra	Yes, if egress records both tracks	Opt-in: configure egress for two tracks	None from Hotato
Pipecat	No: the `setup` recorder runs in your pipeline	Yes, the scaffold writes caller ch0 / agent ch1	Yes, once the recorder is added	None: local recorder, no network call

Automatic pull: does hotato capture fetch it for you. Separate channels: is caller/agent separation what you get. Contract-grade default: is that separation on without extra setup. Network use: what capture calls out to, not your infrastructure.

1. Add the recorder

setup prints an AudioBufferProcessor recorder: input (caller), output (agent), writing one WAV with caller on channel 0.

bash

uvx hotato setup --stack pipecat

Same call, two ways to record it. Two channel: the scorer can tell which party owns the overlap. Mono-mixed: caller and agent are summed onto one track before scoring, so the overlap can’t be attributed to either party.

2. Score the recording

Point Hotato at the WAV the recorder wrote. --expect yield/hold = interruption vs. backchannel.

bash

uvx hotato capture --stack pipecat --stereo captured.wav --expect yield

Scoring more than one call?

Pipecat has no vendor recording API to poll, so pull and sweep have nothing to fetch here. Point hotato analyze at the folder your recorder writes to instead: same ranking sweep produces, over every WAV inside.

bash

uvx hotato analyze ./recordings

The dual-channel gate

Separated caller and agent audio is what makes overlap attributable; speech activity comes from exposed VAD thresholds, frame-inspectable. The scorer takes one two-channel WAV, or two aligned mono WAVs. A single mixed-down track has no channel to attribute overlap to, and unlike a vendor recording there is no fetch step to retry in degraded mode: keep caller and agent on separate channels straight out of the recorder (step 1), or the call is not scorable.