Beyond Text: How Okidoki Understands Photos, Voice Notes & PDFs on WhatsApp

On WhatsApp, real customers don't send tidy paragraphs. They snap a photo of the problem, record a quick voice note, or forward a PDF and expect you to "just get it." Most chatbots choke on anything that isn't text. Okidoki understands the media people actually send — images, audio, documents, even short videos — and feeds all of it into the same AI agent that answers and books for you. That's what makes it one of the most versatile agents on the market.

Most bots read text. Okidoki understands media.

When a message arrives with an attachment, okidoki doesn't ignore it or reply "please describe your issue." It analyzes the attachment, turns it into something the AI can reason about, and continues the conversation as if the customer had explained everything perfectly.

What okidoki can understand

• Photos & screenshots — an image is analyzed with vision AI, so the agent can describe what it sees (a leaking pipe, a broken part, a receipt, a screenshot of an error) and respond to it.
• Voice notes — audio is transcribed automatically, so a rushed 15-second recording becomes text the agent understands and answers.
• PDFs & DOCX documents — files like warranties, quotes, contracts or forms are read and their content extracted, so the agent can pull out the details that matter.
• Short videos — a brief clip is analyzed too (rate-limited to keep things fast), so a customer can show the issue instead of describing it.

How it fits the conversation

The magic isn't just recognizing a file — it's using it. Once the media is understood, its content is injected into the same agent context, so okidoki can answer the real question, check your catalog, or book an appointment based on what the customer sent. A photo of a leak plus a PDF warranty becomes "your repair is covered — here are two visit times." No forms, no back-and-forth.

Customers send what they have. Okidoki meets them there — a photo, a voice note, a document — and turns it into an answer and a next step.

Beyond WhatsApp: the same brain on email

This isn't a WhatsApp-only trick. The same attachment pipeline also powers email on the Growth plan and above — a customer can email a document and get the same understanding and follow-up. One agent, one brain, across your channels.

Honest limits worth knowing

We'd rather be precise than overpromise:

• The AI replies in text on WhatsApp. It understands rich media coming in; when it shares products, photos or brochures back, it sends them as tap-to-open links rather than native media messages.
• The website widget supports text and live voice, but not visitor file uploads — send-us-a-file scenarios shine on WhatsApp and email.
• WhatsApp connects by scanning a QR code from your dashboard, and attachments have a size limit (around 20 MB), with video intentionally rate-limited so replies stay fast.

Why it matters

Every attachment a customer has to explain in words is friction — and a chance to drop off. By understanding photos, voice notes, PDFs and videos, okidoki removes that friction and keeps the conversation moving toward an answer, a reservation, or a booked visit.

Where to go next

See how to connect the channel in how to integrate WhatsApp with website chat, or explore the full agent capabilities on the okidoki developers page.