Ears · The audio sense

Audel

Machine-graded audio feedback a coding agent consumes to hear its own output — and self-correct before it claims done.
$ pip install audel
What it does

Play → hear → report → fix

Hear & report

A machine verdict (pass/warn/fail) with time-grounded issues — loudness (EBU R128), clipping, silence, dropouts, A/V desync. DSP-grounded, no key needed.

Match intent

Grade what the audio says against a brief — language, required phrases, duration — via local faster-whisper ASR.

Listen & critique

analyze adds acoustic + LLM judgment for subjective claims. Egress happens only here; check stays offline by construction.

Streaming & temporal

watch grades behaviour over time; StreamMonitor grades live PCM chunks — playthrough, dropouts, liveness — bounded by construction.

Quickstart

One verdict away

$ pip install audel && audel demo

audel demo grades a deliberately silent render, prints a FAIL report (missing audio — no LLM key), then loops against the fixed version and prints "what changed: issues resolved → PASS."

One ecosystem

Eyes, ears & brain

pip install "verel[hearing]" pulls the whole stack. All three share one contract, agentsensory.