Media Library #

The Media Library lets your AI bot send images, videos, GIFs, voice notes, and PDF documents to a contact during a conversation. You upload the files once, tell the bot when each one should be sent, and the bot picks the right one at the right moment - for example, sending a product photo when someone asks “can I see it?”, a PDF brochure when someone asks for the full price list, or a quick voice note that walks a customer through your menu.

Think of it as a shelf of ready-to-send files for your bot. You decide what’s on the shelf and when each item gets used.

The Media Library is available on every account by default.

Why Use a Media Library #

Text-only answers can only go so far. A photo, short video, or voice note often answers a question faster and more convincingly than a paragraph of text.

What a media library lets your bot do:

• Send a product photo when a contact asks to see something
• Share a menu, price list, or product shot as an image
• Send a short demo or how-to video
• Reply with a GIF when it fits the tone of the conversation
• Send a voice note or audio clip for a personal, human touch
• Send a PDF document - a brochure, price list, catalog, or terms document

The bot only ever sends files you uploaded and approved - it never pulls images from the internet or makes them up.

Where to Find It #

In the main sidebar, under AI Studio, click Media Library. This page shows every file the bot can send in one account-wide view, grouped so a file used in several places appears as a single card. Each item belongs to one or more of your AI Agents — the card lists which. Use the Show filter at the top to view everything or just one agent’s media.

On older accounts that still use classic Campaigns, the same page works identically — items belong to campaigns instead of agents, and the filter lists campaigns.

How the Bot Decides What to Send #

This is the most important part to get right. For each item you upload, you write a short instruction describing when the bot should send it. The bot reads these instructions and only sends an item when the situation you described actually comes up.

For example:

The clearer your instruction, the better the bot’s timing. Vague instructions like “send when relevant” give the bot less to work with than “send when a customer asks about pricing for the premium plan.”

How the Bot Knows What’s in Each File #

When you upload a file, the platform automatically analyses it and writes a short description of what it contains - it “looks at” your image or video, or transcribes your audio. The bot uses this together with your “when to send” instruction, so it understands both what the file is and when to use it - even if your own description is brief. You’ll see this AI summary on each item in the list (shown as “AI sees: …”).

You don’t have to do anything for this - it happens on upload. It just means a clear title and instruction, plus the AI’s own read of the file, give the bot the best chance of sending the right thing at the right moment.

Media in Follow-Ups #

One of the most useful places for media is a follow-up - the automatic nudge the bot sends when a contact who replied earlier goes quiet for a while. A short voice note, a quick video, or a “still there?” clip often re-engages someone far better than another line of text.

To use it, write a “when to send” instruction that describes a follow-up situation, for example:

When that moment arrives, the bot decides whether one of your media items fits the follow-up and, if so, sends it.

Follow-ups can be media-only. If an item is meant to stand on its own - say a voice note that already says everything - the follow-up can be just that media, with no text. Signal this in the “What should the bot say when sending this?” field, e.g. “just send the voice note, no text needed”. Otherwise the bot sends a short text nudge alongside the media.

A few things to know:

• This works for both the quick follow-ups (the early nudges) and the longer multi-day follow-up cycles - on WhatsApp Web, Instagram, Messenger, and Telegram.
• On the WhatsApp Business API, media follow-ups can only go out while the 24-hour window is still open. After that, WhatsApp allows only pre-approved templates, so later follow-ups there stay text-only.
• The same per-conversation send cap applies, so the bot won’t repeat the same file across follow-ups unless you raise the limit on that item.

How to Add Media #

1. In the main sidebar, click AI Studio.
2. Click Media Library.
3. Click the + Add media button.

4. Choose which agent(s) the item should belong to — an item can be attached to more than one. (On older campaign-based accounts you pick campaigns here instead.)
5. Click to pick a file (image, video, GIF, audio, or PDF) from your device. For audio you can also switch to Record voice note and record one right in the app instead of uploading. If you add audio, you can choose whether it’s sent as a voice note (a recorded-voice bubble) or a regular audio file.
6. Type a short Title so you can recognize the item later (e.g. “Blue dress photo”).
7. In the When should the bot send this? field, describe the situation in plain language.

8. Click Save. The item appears in your media list and is immediately available to the bot on every agent (or campaign) you attached it to.

You can also add media from an Agent’s own Media tab — that opens the same underlying library, just pre-filtered to that agent. The tab works from the moment the agent exists, including one you created a minute ago or duplicated from another, and a file you upload there belongs to that agent straight away. On older accounts, a campaign that still holds its own AI settings directly has the same Media step built right into its own editor:

Already uploaded a file and want to reuse it on another agent? Instead of uploading it again, open the item and click Assign, then pick the additional agent(s). The file itself isn’t duplicated — you’re just adding another agent to the same media item.

What to fill in #

These fields are notes to the bot, not messages to your contact #

The Title, When should the bot send this? and What should the bot say when sending this? fields are yours to write in whatever language you think in. Your contact never sees them.

The bot always writes to each contact in that contact’s language. So if you write your notes in Chinese and an English-speaking contact asks to see the venue, the bot sends the photo with an English caption based on your note. Write your notes in the language that is easiest for you.

Supported files and limits #

• Images: JPG, PNG, GIF, WebP
• Video: MP4, MOV, WebM
• Audio: MP3, OGG, WAV, M4A (send as a voice note or a plain audio file)
• Documents: PDF
• Maximum size: 10 MB per file
• Per agent (or campaign): up to 50 items

When you upload a PDF, the platform reads its text so the bot knows what’s inside it - same as the automatic analysis for images and audio.

Editing and Removing Media #

• To change an item’s title or its when-to-send instruction, click the item and edit the fields, then save. (Editing the instruction does not require re-uploading the file.)
• To remove an item, click the delete (trash) icon on it. The file and its instruction are removed and the bot stops sending it.
• To replace the file itself, delete the item and add a new one.

Which Channels Can Receive Media #

Most channels can receive media, but not all:

A couple of things to keep in mind:

• SMS cannot receive media. In SMS conversations the bot simply won’t offer media items.
• Instagram can’t receive documents. On an Instagram conversation the bot won’t send PDFs - it answers in text instead.
• GIFs are automatically optimized when you upload them so they animate on WhatsApp. On some other channels a GIF may show as a still image.
• WhatsApp’s 24-hour window applies to media too. If too much time has passed since the contact last messaged you, WhatsApp may block any message - including media - until they reply again.

Tips for Best Results #

1. Write the “when to send” instruction like you’re briefing a new team member. Describe the exact moment - “when a customer asks to see photos of the apartment” - not just the topic.
2. One purpose per item. Upload separate items for separate situations rather than one file you hope covers everything.
3. Keep files small and fast to load. A compressed image or short clip arrives faster and looks better in chat than a huge file.
4. Use clear, recognizable titles. When your library grows, good titles make it easy to find and update the right item.
5. Test in Try Out. After adding media, use the Agent’s Try Out tab to check the bot picks the right item at the right time before going live. The image, video, audio, or document shows up right in the test chat, exactly where the contact would receive it.

Older test conversations may show “Media preview expired”. The test chat is a sandbox - no real message leaves your account - and its media previews stay viewable for 7 days. If you come back to an old test conversation later, an expired preview shows a small note instead; just ask the bot again to see it fresh. Real conversations are unaffected: contacts receive the actual file, which doesn’t expire.

The bot won’t spam the same file. Each item is sent at most once per conversation by default. If a customer keeps asking about the same thing, the bot answers in text rather than firing the same image again. Raise the “Max times to send per conversation” value on an item if you genuinely want it resendable.

Sending Media Yourself #

The Media Library is for the bot’s automatic sends. You can always send any image, video, or file to a contact manually from the Chat Interface at any time - that doesn’t require setting up a library item.

Next Steps #

• AI Agents — the Media tab inside each Agent’s editor.
• AI Agents — the main page of the AI Studio group, alongside Knowledge Base and Custom Functions.
• Automated Follow-Ups — the full guide to the follow-up sequences media can attach to.