MimicScribe
← Back to Bugs

I was recording a Zoom meeting, left the meeting and MimicScribe crashed

Closed

#49 · by User #942f · Jul 15, 2026 · Released v1.0.0-rc.18

I was recording a Zoom meeting, left the meeting and MimicScribe crashed. On restart, it told me it was recovered and let me open finder to the directory. However, on trying to import the two files, neither will import.

mimicscribe-system-184FFF41-F139-4E2F-9E20-100F05A5E885.wav: No audio track found in file mimicscribe-mic-184FFF41-F139-4E2F-9E20-100F05A5E885.wav: No audio track found in file

They're too big, and also contain sensitive info, so I can't share them. This is the only log entry in the system tap dropout log:

2026-07-15T19:02:18Z holeMs=16725 mode=SILENT BUFFERS (callbacks firing on schedule, all-zeros audio) aggDevice=155 running=1 nominalRate=48000 defaultOutputDevice=103(john’s AirPods 4) [no output switch this session]

Also, FWIW, quicktime crashes: The document “mimicscribe-system-184FFF41-F139-4E2F-9E20-100F05A5E885.wav” could not be opened. An unknown error occurred (-12842)

Comments (4)

User #6f81 · Jul 16, 2026, 12:34 AM

Ah I'm sorry. I'm looking into a fix now. Thanks for the details

User #6f81 · Jul 16, 2026, 12:37 AM

Apparently the WAV files headers aren't getting finalized. Here's a workaround to get the them playing hopefully:


Their two files can be fixed in place with a dependency-free Python one-liner (macOS ships python3). Safe: it only rewrites the two size fields, touching no audio bytes — though suggesting they duplicate the files first is prudent:

python3 -c '
import struct,sys
for p in sys.argv[1:]:
    f=open(p,"r+b"); f.seek(0,2); size=f.tell(); f.seek(12)
    while True:
        h=f.read(8)
        if len(h)<8: break
        tag,csz=h[:4],struct.unpack("<I",h[4:])[0]
        if tag==b"data":
            off=f.tell()
            if csz!=size-off:
                f.seek(off-4); f.write(struct.pack("<I",size-off))
                f.seek(4); f.write(struct.pack("<I",size-8))
                print(p,"repaired:",size-off,"bytes of audio")
            else: print(p,"already OK")
            break
        f.seek(csz+(csz&1),1)
    f.close()
' mimicscribe-mic-*.wav mimicscribe-system-*.wav```
User #6f81 · Jul 16, 2026, 2:48 AM

Sent an email with more info. Will have the second half of this fix in the next release. (If we root caused it correctly 🤞). Thanks for reporting!

Marshall (developer) · Jul 19, 2026, 10:07 PM

This shipped in v1.0.0-rc.18. Update via Preferences → Check for Updates, or download from mimicscribe.app, then try importing those two files again.

Also new in this release: a "My mic is always me" toggle in Settings that treats your microphone channel as a single speaker and skips diarization on it — useful if your own audio was ever getting split across multiple speakers.

0/2000