TechLatest

Posted on Feb 18 • Originally published at Medium

From Chat App to AI Powerhouse: Telegram + OpenClaw

#openclaw #opensource #ai #llm

In this guide, we’ll walk through how to install OpenClaw on macOS and connect it to Telegram step by step.

OpenClaw is a self-hosted AI assistant that runs locally on your machine while connecting to messaging platforms like Telegram, WhatsApp, Discord, and more. Unlike typical cloud chatbots, OpenClaw gives you full control over your model, workspace, sessions, and integrations.

By the end of this tutorial, you will have:

OpenClaw installed on macOS
An AI model configured (using OpenAI in this example)
Telegram bot connected and approved
Gateway service running in the background
Web dashboard and TUI are working
A fully functioning Telegram AI assistant Let’s get started.

Step 1: Install OpenClaw on macOS

Before connecting OpenClaw to Telegram, we first need to install it locally.

OpenClaw provides a modern installer that automatically detects your operating system and installs the required dependencies (including Node.js if needed).

Run the Installer

Open your terminal and run:

curl -fsSL https://openclaw.ai/install.sh | bash

This command will:

Detect your operating system
Verify Homebrew (if on macOS)
Check Node.js installation
Install OpenClaw globally via npm
Prepare your environment
Installation Output

If everything is configured correctly, you should see output similar to this:

OpenClaw Installer
Detected: macOS
Install method: npm
Requested version: latest

[1/3] Preparing environment
✓ Homebrew already installed
✓ Node.js v24.6.0 found

[2/3] Installing OpenClaw
✓ Git already installed
INFO Installing OpenClaw v2026.2.15
✓ OpenClaw npm package installed
✓ OpenClaw installed

[3/3] Finalizing setup

OpenClaw installed successfully (2026.2.15)!
Installation complete. Your productivity is about to get weird.

INFO Starting setup
OpenClaw 2026.2.15

This confirms:

Your dependencies are ready
The latest OpenClaw version is installed
Setup has successfully initialized

That means OpenClaw has successfully identified your environment and is preparing the installation.

Step 2: Complete Onboarding & Review Security Warning

After installation, OpenClaw automatically launches its onboarding process.

You’ll be greeted with the OpenClaw interface along with an important Security Warning screen.

Why This Warning Matters

OpenClaw is not a simple chatbot. It is an autonomous AI assistant capable of:

Reading local files
Executing system actions
Running tools and automations
Connecting to external services

Because of this, OpenClaw clearly warns users that:

A bad prompt can trick it into doing unsafe things.

This is expected behavior for agent-style systems.

Step 3: Configure the Model & Authentication Provider

After completing onboarding, OpenClaw will prompt you to select a Model/Auth Provider.

This is where you decide which LLM backend will power your assistant.

During setup, you’ll see a list of available providers. OpenClaw supports multiple options, allowing you to choose based on performance, pricing, availability, or infrastructure preference.

Supported Model Providers

OpenClaw provides flexibility by supporting a wide range of AI providers, including:

Anthropic
OpenAI
OpenRouter
LiteLLM
Amazon Bedrock
Vercel AI Gateway
Moonshot AI
MiniMax
OpenCode Zen
GLM Models
Z.AI
Synthetic
Qianfan

This allows you to run OpenClaw using:

Direct API providers
Gateway aggregators
Enterprise cloud backends
Local or proxy-based model routing

You can choose whichever provider best fits your workflow and infrastructure.

Example Setup: OpenAI

In this tutorial, we’ll use OpenAI as the model provider.

During configuration:

Select OpenAI from the list
Choose the authentication method:
OpenAI API Key
Enter your API key when prompted
You’ll see a prompt similar to:

Enter OpenAI API key

After entering a valid key, OpenClaw verifies the connection and continues setup.

Step 4: Select and Confirm the Default Model

After entering your API key, OpenClaw verifies it and automatically configures a default model.

You’ll see confirmation like:

Saved OPENAI_API_KEY to ~/.openclaw/.env
Model configured
Default model set to openai/gpt-5.1-codex

This means:

Your API key is securely stored in the local .env file
The model provider connection is successful
A default model has been selected

Choosing Your Default Model

OpenClaw gives you the flexibility to:

Keep the current default model
Enter a custom model manually
Select from the available models provided by your chosen backend

For example, with OpenAI, you may see options like:

openai/gpt-5.1-codex
openai/codex-mini-latest
openai/gpt-4
openai/gpt-4-turbo
openai/gpt-4.1
openai/gpt-4.1-mini
openai/gpt-4o and more…

For this tutorial, we’ll use gpt-4o:

openai/gpt-4o

Step 5: Connect OpenClaw to Telegram

After configuring your model, OpenClaw will ask you to select a chat channel.

You’ll see something like:

Select channel (QuickStart)

● Telegram (Bot API)
○ WhatsApp (QR link)
○ Discord (Bot API)
○ Slack (Socket Mode)
...

Step 6: Enter Your Telegram Bot Token

After selecting Telegram (Bot API) as your channel, OpenClaw will prompt you to enter your Telegram bot token.

You’ll see instructions like:

Telegram bot token

1) Open Telegram and chat with @BotFather
2) Run /newbot (or /mybots)
3) Copy the token (looks like 123456:ABC...)

Generate the Token (If You Haven’t Already)

Open Telegram
Search for @botfather
Run:

/newbot

Choose a bot name
Choose a username (must end in bot)

BotFather will return a token similar to:

123456789:AAExampleGeneratedToken

Enter the Token in OpenClaw

Back in your terminal, paste the token when prompted:

Enter Telegram bot token

OpenClaw will validate the token and complete the channel configuration.

If successful, Telegram is now linked to your OpenClaw instance.

Optional: Set via Environment Variable

Instead of pasting it interactively, you can also set it manually:

export TELEGRAM_BOT_TOKEN=your_token_here

Or add it to:

~/.openclaw/.env

This is useful for production or automated deployments.

Step 7: Channel Confirmed & Workspace Initialized

After entering your Telegram bot token, OpenClaw finalizes the channel configuration.

You’ll see confirmation like:

Selected channels

Telegram — simplest way to get started — register a bot with @BotFather and get going.

Updated ~/.openclaw/openclaw.json
Workspace OK: ~/.openclaw/workspace
Sessions OK: ~/.openclaw/agents/main/sessions

This confirms:

Telegram channel is successfully configured
The configuration file has been updated
Workspace directory is initialized
Agent session storage is ready At this point, your assistant is fully wired to Telegram.

What Just Happened?

OpenClaw created and validated:

~/.openclaw/openclaw.json → Main configuration
~/.openclaw/workspace → Working directory
~/.openclaw/agents/main/sessions → Conversation memory & state

This is where your agent:

Stores sessions
Maintains chat history
Tracks tool execution
Manages runtime state
Everything is now persistent and structured.

Skills Status Overview

You’ll also see a skills report:

Skills status

Eligible: 7
Missing requirements: 42
Unsupported on this OS: 0
Blocked by allowlist: 0

What This Means

Eligible → Skills ready to run
Missing requirements → Optional skills requiring extra dependencies
Unsupported on this OS → OS-incompatible tools
Blocked by allowlist → Restricted skills This doesn’t mean something is broken — it simply reflects optional capabilities not yet installed.

You can install additional skills later as needed.

Step 8: Gateway Service Installed & Telegram Connected

After completing the configuration, OpenClaw installs and starts the Gateway service.

You’ll see output similar to:

Installing Gateway service....
Gateway service installed.

Telegram: ok (@yourbotname)
Agents: main (default)
Heartbeat interval: 30m
Session store: ~/.openclaw/agents/main/sessions/sessions.json

What Just Happened?

OpenClaw has now:

Installed the Gateway service (runs in the background)
Created a macOS LaunchAgent
Linked your Telegram bot successfully
Initialized the default agent
Set up persistent session storage The important line is:

Telegram: ok

That means your bot is live.

What is the Gateway?

The Gateway is the control plane of OpenClaw.

It:

Manages chat channels
Handles agent sessions
Maintains WebSocket connections
Routes messages between Telegram and your model
Stores session history
Runs continuously in the background

On macOS, it installs as a LaunchAgent:

~/Library/LaunchAgents/ai.openclaw.gateway.plist

Logs are stored at:

~/.openclaw/logs/gateway.log

Step 9: Hatch the Bot in TUI

When you see:

How do you want to hatch your bot?
Hatch in TUI (recommended)

Do this:

Just press Enter.

That runs:

openclaw tui --ws://127.0.0.1:18789 --agent main --session main

You’ll then see:

connected | idle
agent main | session main
openai/gpt-4o

That means:

Gateway running ✅
Model connected ✅
Session active ✅
Bot alive ✅

You can now type:

Hello

and it responds.

Step 10: Open Dashboard & Start Chatting

Now your Gateway is running.

You can open the OpenClaw WebUI (Dashboard) in your browser:

http://127.0.0.1:18789

If a token was generated during setup, open it like this:

http://127.0.0.1:18789/#token=YOUR_TOKEN

(Use the exact token shown in your terminal.)

What you should see:

Status: Connected
Health: OK
Chat interface ready

To Chat

Go to Chat (left sidebar)
Type your message in the input box
Press Enter or click Send Example test message:

What is 12 × 14?

If it replies → your model, gateway, and agent are fully working.

Step 11: OpenClaw Access Configuration

Run this command to approve Telegram access:

openclaw pairing approve telegram your_code

Step 12: Chat with Telegram Bot

Open Telegram and send a message (e.g., "hello") to your bot.

If it replies, your Telegram bot is successfully connected and working.

Conclusion

You now have a fully working OpenClaw setup connected to Telegram.

In this walkthrough, we:

Installed OpenClaw on macOS
Configured the model provider and API key
Selected a default model
Connected Telegram using BotFather
Installed and started the Gateway service
Launched the TUI and Web Dashboard
Approved Telegram pairing
Successfully tested the bot

At this point, your assistant is live and running locally on your machine. Messages sent to your Telegram bot are routed through the OpenClaw Gateway, processed by your selected model, and returned in real time.

From here, you can explore:

Adding more skills
Connecting additional channels (WhatsApp, Discord, Slack)
Enabling automation workflows
Integrating tools like Notion, GitHub, or file access
Running deeper security audits

OpenClaw turns your local machine into a personal AI control center. Now it’s your move — start building, automating, and experimenting.

Note

We’re excited to offer an exclusive opportunity for our community. You are eligible to receive free Azure credits worth $1,000 to explore, test, and deploy any TechLatest product or solution. Whether you’re building AI agents, experimenting with automation tools, or scaling your infrastructure, these credits give you the flexibility to innovate without upfront costs. This is a limited-time offer available on a first-come, first-served basis, so don’t miss out. To claim your credits, simply fill out the form here: https://www.techlatest.net/free_azure_credits/ and explore our available products and solutions here: https://techlatest.net/support/.