Telegram Bot Commands¶

/start¶

Lists all public presets (where private: false) as inline keyboard buttons. Each button shows the preset's display name.

Also triggered by /presets and by sending any text message when not in the middle of a flow.

After selecting a preset, the bot guides the user through a conversation:

User taps a preset button. The bot shows:

If the preset's prompts contain placeholders, the bot asks each question in sequence:

Choice placeholders ([[question:opt1|opt2]]) -- shown as inline keyboard buttons
Free text placeholders ([[question]]) -- the bot asks the user to type their answer

Scene-based placeholders include the scene number (e.g., "Scena 1: style").

The bot asks the user to send a photo. The photo is used as the input image for the preset's workflow.

Note

The bot always asks for a photo. For text-to-image presets that do not require an input image, the photo is sent but the workflow ignores it.

After receiving the photo, the bot shows a confirmation message with two buttons:

After the user confirms:

The bot sends the preset to POST /api/admin/presets/{id}/run with the photo and placeholder answers
The bot displays "Uploading and queuing job..."
Every 10 seconds, the bot polls GET /api/run/status/{prompt_id}:
Queued -- "In coda -- waiting for execution slot..."
Running -- shows percentage, current node, and ETA
Loading models -- "Loading models..." (0% with no node)
On completion, the bot downloads the output file and sends it to the chat:
Video files (.mp4, .webm) are sent as Telegram video messages
Image files are sent as Telegram photo messages
The caption includes the preset name and seed value