Documentation Index
Fetch the complete documentation index at: https://mintlify.com/AmiraliNotFound/dummy-gemini-bot/llms.txt
Use this file to discover all available pages before exploring further.
The Settings tab exposes all keys from the config table in the SQLite database as editable fields. Changes are saved via POST /api/config, which updates both the persistent database record and the in-memory runtime_config dictionary immediately — no bot restart is required.
Configuration Fields
AI Model
| Key | Description |
|---|
MODEL_ID | Primary Gemini model used for responses (e.g. gemini-2.5-flash) |
FALLBACK_MODELS | Comma-separated list of fallback models tried in order when the primary fails (e.g. gemini-2.5-flash-lite,gemini-2.5-flash,gemma-4-31b-it) |
CONTEXT_LIMIT | Number of recent messages included in the AI context window |
TIMEOUT | Request timeout in seconds for Gemini API calls |
TTS
| Key | Description |
|---|
TTS_ENGINE | Active TTS backend — edge (Microsoft Edge TTS) or gemini |
TTS_GEMINI_MODEL | Comma-separated Gemini TTS model(s) (e.g. gemini-2.5-flash-preview-tts) |
TTS_GEMINI_VOICE | Voice name for the Gemini TTS engine (e.g. Kore) |
TTS_EDGE_VOICE | Voice name for the Edge TTS engine (e.g. fa-IR-FaridNeural) |
TTS_FALLBACK_TO_EDGE | True/False — fall back to Edge TTS if Gemini TTS fails |
TTS_VOICE_PITCH | Pitch multiplier for TTS output (float, default 1.0) |
Behavior
| Key | Description |
|---|
SYSTEM_INSTRUCTION | Global system prompt that defines the bot’s default persona |
PERSONA_PRESETS | JSON array of preset persona objects — each with name and prompt fields |
RANDOM_ROAST_CHANCE | Probability (float 0.0–1.0) that the bot spontaneously roasts a message without being directly addressed |
Discussion Mode
| Key | Description |
|---|
DISCUSSION_MODE_ENABLED | True/False — enable scientific/technical discussion mode |
DISCUSSION_PROMPT | System instruction override used when discussion mode is active |
Daily Summaries
| Key | Description |
|---|
DAILY_SUMMARY_ENABLED | True/False — enable automatic daily chat summaries |
DAILY_SUMMARY_TIME | Time of day to send the summary (24-hour format, e.g. 00:00) |
DAILY_SUMMARY_PROMPT | Prompt used to generate the daily summary |
Rate Monitoring
| Key | Default | Description |
|---|
MONITOR_LIMIT_RPM | 15 | RPM ceiling for text generation models (used in Limits tab progress bars) |
MONITOR_LIMIT_RPD | 1500 | RPD ceiling for text generation models |
MONITOR_LIMIT_TTS_RPM | 15 | RPM ceiling for TTS models |
MONITOR_LIMIT_TTS_RPD | 1500 | RPD ceiling for TTS models |
Saving Changes
Edit any field in the Settings tab and click Save. The dashboard sends a POST /api/config request with the updated key-value pairs:
POST /api/config
Authorization: Bearer <Telegram WebApp initData>
Content-Type: application/json
{
"MODEL_ID": "gemini-2.5-flash",
"RANDOM_ROAST_CHANCE": "0.05"
}
config SQLite table via save_config_key and immediately updates the live runtime_config dictionary, so the bot picks up the change on the very next incoming message.
The Settings tab also exposes two maintenance utilities for the media downloader:
Upload Cookies
POST /api/upload_cookies — Upload a fresh cookies.txt file for yt-dlp authenticated downloads. Paste the contents of your Netscape-format cookie export into the upload field and submit. The file is saved to data/cookies.txt (and synced to the root cookies.txt as a fallback).
Update yt-dlp
POST /api/update_ytdlp — Triggers a pip install --upgrade yt-dlp subprocess on the server to update the scraper to the latest published version. The response includes the pip output log. No SSH access or server restart is needed.
Keep yt-dlp updated regularly — Instagram, YouTube, and Threads frequently change their internal APIs and anti-bot protection mechanisms, which can break media downloads until the scraper is patched.
SYSTEM_INSTRUCTION and DISCUSSION_PROMPT support long multi-line text and can be edited in the full-size textarea in the Settings tab. PERSONA_PRESETS must be a valid JSON array where each element is an object with "name" and "prompt" string fields. Invalid JSON will cause preset-related features to fall back to defaults.
