Setup Guide
Detailed walkthrough for configuring your Telegram bot and agent.
Creating your Telegram bot
Step 1: Talk to BotFather
Open Telegram and search for @BotFather (the official Telegram bot for managing bots). Send it the command:
BotFather will ask you two things:
- A display name for your bot (e.g., "My AI Assistant")
- A username that ends in
bot(e.g.,my_ai_assistant_bot)
Step 2: Copy your bot token
After creating the bot, BotFather sends you an API token. It looks something like:
Keep this token private. Anyone with it can control your bot.
Step 3: Optional bot settings
While you're in BotFather, you can optionally:
/setdescription— set what users see before starting a chat/setabouttext— set the bio shown on the bot's profile/setuserpic— upload a profile photo for your bot
These are cosmetic but make the experience feel more polished.
Connecting to ClawLuv
Paste your token
On the ClawLuv dashboard, find the Bot Token field and paste the token from BotFather. Click Deploy.
ClawLuv validates the token, provisions your container, and connects to Telegram. This typically takes 30–60 seconds.
Verify it's working
Open your bot in Telegram and send a simple message like "Hello". Your agent should respond within a few seconds.
If it doesn't respond, check Troubleshooting.
Understanding your agent
Memory
Your agent uses pgvector to store and retrieve conversation history. This means it can recall things you mentioned days or weeks ago, as long as they're relevant to the current conversation.
Memory is scoped to your container — nobody else can access it.
Skills
Skills are modular capabilities your agent can use. Default skills include:
| Skill | What it does |
|---|---|
| Web search | Searches the web and summarizes results |
| Voice transcription | Converts Telegram voice notes to text using Whisper |
| Image analysis | Describes and reasons about photos you send |
| Self-improvement | Logs errors and corrections to get better over time |
You can install additional skills from the skills marketplace on your dashboard.
Self-improving loop
When you correct your agent ("No, I meant X" or "That's wrong, it should be Y"), it logs the correction and adjusts its behavior for next time. This happens automatically — no configuration needed.
Configuration options
Most users don't need to change anything beyond the bot token. But if you want to customize:
- Model selection — choose which AI model your agent uses (defaults to Claude)
- System prompt — customize your agent's personality and instructions
- Skills — enable or disable specific capabilities
All configuration is done through the ClawLuv dashboard. No files to edit, no CLI commands to run.