Multimedia interpretation

Last updated: August 5, 2026

Leads rarely stick to plain text. They send screenshots, photos, PDFs, voice notes and short videos. Multimedia Interpretation lets Charlie receive that media, understand what is in it, and continue the conversation using that information.

How it works

1. Ingestion.

When a lead sends a message with an attachment, Charlie detects it, classifies the media type (image, file, audio, video), normalizes it into one internal format regardless of the platform it came from, and stores it securely with metadata (media type, format, size, source platform, timestamp, lead and conversation reference, processing status).

2. Interpretation. The right engine processes the media:

   - Images and files (Priority 1): a vision model extracts text and visual context from images; PDFs and documents are parsed for text, with OCR applied to scanned or image-based pages.

   - Audio (Priority 2): voice notes are transcribed to text by a speech-to-text engine, with format normalization first.

   - Video (Priority 3): the audio track is transcribed and visual keyframes are analyzed, then merged into one interpretation. If only one track exists (silent screen recording, or audio over a black screen), Charlie processes what is available and marks which tracks were present.

3. Use in the conversation. The interpretation is added to the conversation context and used by Charlie's features (supported for the Qualification and Autobooks flows) exactly as if the lead had typed it.

Benefits

  • Smoother bookings. Screenshots of availability or a voice note confirming a time are understood and used to continue the booking flow.

  • Lower human workload. Fewer conversations need a human to open the attachment and interpret it manually.

  • Human like conversation

Configuration

Enabling it

1. Open the right-side Settings menu and go to the General section.

2. Turn on the Multimedia Support toggle. It is disabled by default on existing workspaces.

3. The first time it is enabled, a one-time pop-up explains credit consumption and asks for consent — accept it to activate.

Credit consumption:

- Image or file: 3 credits per file

- Audio: 5 credits per minute

- Video: 8 credits per minute

Audio and video are pro-rated and rounded up to the nearest whole credit (a 30-second voice note = 3 credits; a 30-second video = 4 credits).

Usage shows up in your DM consumption, and appears in transaction history as "Multimedia interpretation". Auto-recharge, if enabled, covers it automatically. Your billing page shows one balance with a breakdown of messaging versus multimedia usage for the cycle.

There is no separate multimedia purchase. Multimedia draws from the same credit balance as your messaging

Practical tips

- Test on the channel your client actually uses — each platform compresses and formats media differently.

- Watch credit burn in the first weeks after enabling it for a media-heavy audience, and set a hard daily cap.

- If leads frequently send unreadable photos, add a qualification instruction asking for the specific detail you need in text.

- Interpretation is inbound only. To have Charlie send files, see the "Sending Multimedia Files" article.