Telegram: Send Image Album

Once you've got Telegram: Send Image in one workflow, you want it in all of them - and then you hit the awkward bit: you've got five outputs and you really don't want five separate messages spamming the chat, one per render. Telegram: Send Image Album is the tidy version. It collects up to ten image sources and drops them as a single Telegram album (sendMediaGroup), so your grid of variations arrives as one swipeable message instead of a wall of notifications.

It's the same pack, same credentials, same idea - the node exists to replace a dozen Telegram: Send Image nodes with one. If your workflow fans out into several outputs (different seeds, different upscalers, a before/after pair), this is the one you reach for.

How it works

It's a straightforward gather-and-send. Each of the optional image_1 through image_10 inputs takes an IMAGE tensor, and any of them may itself be a batch - all frames get flattened in order into one list. Then it calls Telegram's sendMediaGroup, which has a hard cap of 10 media per album, so anything larger is automatically split into multiple albums. If only one image ends up connected, it gracefully degrades to a plain single sendPhoto / sendDocument instead of building a one-item album.

In document mode it embeds the ComfyUI workflow metadata into each PNG, exactly like the single-image node - so the whole album is round-trippable back into ComfyUI. The caption lands on the first item of the first album only; Telegram won't repeat it per media.

The inputs that matter

• image_1 … image_10 (all optional IMAGE) - wire up to ten image sources. Unconnected inputs are simply ignored, so you can leave holes.
• caption (STRING, multiline) - the message text, attached to the first item of the first album.
• send_as_photo (boolean, default True) - sends the collection as a photo album. Preview-friendly, but Telegram recompresses and strips metadata.
• send_as_document (boolean, default False) - sends as a document album of PNGs with the workflow metadata preserved.

No outputs - it's an output node, so it ends the workflow.

Installing it

ComfyUI Manager, search "TelegramSender". Or:

cd ComfyUI/custom_nodes
git clone https://github.com/smbdsbrain/ComfyUI-TelegramSender

In ComfyUI's Python environment:

pip install requests

imageio / imageio-ffmpeg from the pack README are only needed by the video nodes - for images, requests alone is the real dependency. Restart ComfyUI, node under utils/telegram.

Credentials are shared across the pack, first match wins: ComfyUI Settings → Telegram Sender → Credentials (recommended), a telegram_config.json next to the node, or TELEGRAM_BOT_TOKEN / TELEGRAM_CHAT_ID env vars. Bot token from @BotFather; chat ID numeric or @channel.

Troubleshooting

• Got two albums instead of one - both toggles are doing work: send_as_photo and send_as_document each fire their own sendMediaGroup. Keep one. (The defaults - photo on, document off - are actually sane here, unlike the single-image node.)
• More than 10 images - you get multiple albums, split automatically. That's Telegram's rule, not a bug.
• Caption missing from later items - expected. Telegram only shows it on the first media of an album.
• "bot_token and/or chat_id not specified" - the usual pack credential error; check the Settings page.

The video equivalent - same gather-into-album shape, but for VIDEO outputs - is Telegram: Send Video Album.