API reference · Platform

Prompt dictation

Speak a prompt, get the text back — free.

POST/v1/prompts/dictateFree

The mic in every studio prompt box, as an API: send ONE spoken prompt as multipart/form-data — a file part, wav or mp3, up to 60 seconds (2.5MB) — and get its text back. Any language; language is an optional hint.

Free, like prompt enhance, and fenced the same way: one spoken prompt per request and its own rate limit (12 a minute). Longer audio belongs on the paid [transcription](/docs/transcriptions) endpoint.

Request body

filefilerequired

The spoken prompt — wav or mp3, 60 seconds at most.

languagestring

ISO language hint, e.g. en, fr. Omit it to auto-detect.

curl https://eroq.ai/v1/prompts/dictate \
  -H "Authorization: Bearer $EROQ_API_KEY" \
  -F [email protected]
/v1/prompts/dictate
→ POST /v1/prompts/dictate
Press run — this replays a real exchange from the docs' own data. No key, no request, no charge.

Response

200 · application/json
{
  "text": "A woman walking along the beach at sunset, slow tracking shot, warm backlight."
}