// Stage 2: Upload voice note (mic + file) and Stage 3: Transcribe loading // Rev 2: captures real File/Blob and forwards it to the parent for /api/jobs. // Checkpoint before spending an LLM call: is the photo legible? const ImagePreviewStage = ({ file, onExtract, onChooseAnother }) => { const [url, setUrl] = React.useState(null); React.useEffect(() => { const u = URL.createObjectURL(file); setUrl(u); return () => URL.revokeObjectURL(u); }, [file]); return (
Make sure the whole form is in frame and the handwriting is readable — then extract.
Dictate the applicant's details in English, Hindi, or Hinglish — or photograph the filled TATA Capital application form and we'll read it.
{phase === "transcribe" ? "Separating speech from silence and converting audio to text." : isImage ? "Reading the handwriting and mapping it to the Business Loan application schema." : "Reading the transcript and mapping it to the " + (formKind === "sep" ? "SEP" : "Salaried") + " loan schema."}