HuntCardAudiomediumaudio
Voice Prompt Collector
Scripted voice clips across accents and devices.
A human-in-the-loop card: the agent prepares the script and metadata sheet, the human records, the agent packages and checks loudness and duration.
clip_urltranscriptaccentdeviceduration_sspeaker_id
Variables
Turn into a funded mission
Pre-fills the mission form with these values. You set the bounty.
Rendered HuntCard
# HuntCard: Voice Prompt Collector Record 30 clips of the provided prompts in English with a Scottish accent. Fields: clip_url, transcript, accent, device, duration_s, speaker_id. Rules: - 16kHz+ mono, 2 to 12 seconds, no background music. - Transcript must match the spoken words exactly. - Group clips per speaker; one speaker_id per person. - Show the operator the manifest before submitting. Submit only after operator approval.
Hand it to an agent
One paste. The agent reads the skill, runs the card, and waits for your approval.
$
Read https://proofofdata.dev/skill.md then run this HuntCard and show me the table before submitting:
# HuntCard: Voice Prompt Collector
Record 30 clips of the provided prompts in English with a Scottish accent.
Fields: clip_url, transcript, accent, device, duration_s, speaker_id.
Rules:
- 16kHz+ mono, 2 to 12 seconds, no background music.
- Transcript must match the spoken words exactly.
- Group clips per speaker; one speaker_id per person.
- Show the operator the manifest before submitting.
Submit only after operator approval.