OpenClaw India Setup Guide: WhatsApp + Local Models (March 2026)

OpenClaw is free, open-source, and runs entirely on your own hardware. For Indian users, this combination is unusually powerful: it means you get a capable AI agent with zero monthly subscription cost, full data privacy, and the ability to message it from WhatsApp or Telegram — apps already on every Indian smartphone.

This guide walks you through the complete India-specific setup: local models via Ollama, Telegram channel configuration, WhatsApp setup, Hindi prompts, and the India-specific automations that make OpenClaw genuinely useful for daily life on the subcontinent.

Why OpenClaw Is Perfect for India

Three things make OpenClaw stand out for Indian users specifically:

Cost. Most AI tools charge $20-25/month. OpenClaw is free and open-source. Run it with local Ollama models and your total monthly AI spend is ₹0. Even if you connect a cloud API for some tasks, the cost is pay-per-use — not a subscription you forget about.

Privacy. India's DPDP Act 2023 (Digital Personal Data Protection Act) requires careful handling of personal data. Running OpenClaw locally with Ollama means your data never leaves your device. No servers abroad, no third-party processing, no compliance headache.

WhatsApp and Telegram. India has 500 million+ WhatsApp users. The fact that OpenClaw natively supports messaging through WhatsApp and Telegram — apps already on every Indian phone — means you can interact with your AI agent the same way you message family members. No new app to learn, no new habit to build.

Prerequisites

Before starting, make sure you have:

• A computer running Linux, macOS, or Windows (any modern OS works)
• Node.js 22+ installed (or use the Ollama quick-start which handles this)
• 8GB RAM minimum (16GB recommended for local models)
• An internet connection for initial downloads
• A Telegram account (recommended for India setup)
• 30-45 minutes

Optional but recommended:

• A Groq account (free at groq.com) for fast cloud model fallback
• For WhatsApp: A WhatsApp Business account and Meta Developer access

How to Install OpenClaw on Your System

Step 1: Install Node.js 22

# Using nvm (recommended for all platforms)
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.0/install.sh | bash
source ~/.bashrc   # or ~/.zshrc on macOS
nvm install 22
nvm use 22
node --version     # Should show v22.x.x

On Windows, download the Node.js 22 installer from nodejs.org and run it directly.

Step 2: Install OpenClaw

# Install OpenClaw globally via npm
npm install -g openclaw

# Verify installation
openclaw --version

Step 3: Create Your Workspace

# Create and enter a new workspace
openclaw init my-india-agent
cd my-india-agent

Step 4: Run the Setup Wizard

openclaw setup

The setup wizard will ask you:

• Which AI model to use (select "Ollama - Local & Free" for ₹0 setup)
• Which messaging channel to connect (select "Telegram" for easiest India setup)
• Whether to enable memory (select "Yes")

If you prefer the manual setup, continue with the steps below.

Setting Up Telegram (Recommended for India)

Telegram is the recommended messaging channel for Indian OpenClaw users. Setup takes about 5 minutes and requires only a Telegram account — no business registration, no approval process, no fees.

Step 1: Create a Telegram Bot

1. Open Telegram and search for @BotFather
2. Start a conversation and send: /newbot
3. BotFather will ask for a bot name — enter something like My OpenClaw Agent
4. Then it asks for a username — enter something like my_openclaw_bot (must end in bot)
5. BotFather will reply with your bot token — a long string that looks like 7234567890:AAF_abc123...
6. Copy this token — you'll need it in the next step

Step 2: Configure OpenClaw With Your Token

# In your OpenClaw workspace
openclaw config set telegram_token YOUR_BOT_TOKEN_HERE

# Enable Telegram channel
openclaw channel enable telegram

Step 3: Start OpenClaw and Test

openclaw start

# In a separate terminal, or in the logs, you'll see:
# ✓ Telegram bot connected — message @my_openclaw_bot to begin

Open Telegram, search for your bot by username, start a conversation, and send: Hello, what can you do?

You should receive a response within 5-10 seconds. Your OpenClaw Telegram bot is live.

Step 4: Enable Telegram on Mobile

Since your OpenClaw server runs on your laptop or desktop, you need the computer to be on for the bot to respond. For always-on availability, consider running OpenClaw on a low-cost VPS (a ₹800/month Contabo VPS is sufficient) or a Raspberry Pi.

Setting Up WhatsApp (Advanced — Business API Required)

WhatsApp integration requires more setup than Telegram because Meta requires a WhatsApp Business API account. This is not a quick 5-minute setup, but it unlocks the ability to message your OpenClaw agent from the same WhatsApp number you use every day.

Requirements:

• A WhatsApp Business account (free app, available on the Play Store)
• A Meta Developer account (free at developers.facebook.com)
• A dedicated phone number for the bot (a second SIM or a virtual number works)
• Note: Personal WhatsApp numbers cannot be used directly with the Business API

High-Level Setup Steps

1. Create a Meta Developer account and create a new app of type "Business"
2. Add the WhatsApp product to your app
3. Follow Meta's phone number verification process for your business number
4. Generate a permanent access token in the Meta Developer portal
5. Configure OpenClaw:

openclaw config set whatsapp_token YOUR_META_ACCESS_TOKEN
openclaw config set whatsapp_phone_id YOUR_PHONE_NUMBER_ID
openclaw channel enable whatsapp

For most Indian personal users, Telegram is the better choice — it has no approval process, works immediately, and is used by a large and growing share of India's tech community. WhatsApp integration is worth the effort only if you need to use a single number for everything.

Running OpenClaw for Free with Ollama + Llama 3.2

This is the ₹0 setup that most Indian power users run. Ollama lets you download and run open-source LLMs locally on your own hardware.

Step 1: Install Ollama

# Linux/macOS
curl -fsSL https://ollama.com/install.sh | sh

# Windows
# Download the installer from https://ollama.com/download

Step 2: Pull the Recommended Models

The current (March 2026) recommended models for OpenClaw local use:

# Primary model — best quality for general tasks
ollama pull qwen2.5:14b

# Fast model — lightweight for simple tasks
ollama pull llama3.2:3b

# Embedding model — needed for memory search
ollama pull nomic-embed-text

Note: Qwen 2.5 14B requires about 9GB of disk space and 10GB of RAM. If your machine has only 8GB RAM, use qwen2.5:7b (5GB) or llama3.2:3b (2GB) instead.

Step 3: Configure OpenClaw to Use Ollama

# Set Ollama as your model provider
openclaw config set model_provider ollama
openclaw config set model qwen2.5:14b
openclaw config set fast_model llama3.2:3b
openclaw config set embed_model nomic-embed-text
openclaw config set api_url http://localhost:11434

# Verify Ollama is running
ollama list

Step 4: Start OpenClaw With Local Models

# Start Ollama (if not already running as a service)
ollama serve &

# Start OpenClaw
openclaw start

Test by sending a message to your Telegram bot. The first response will be slightly slower as the model loads into memory. Subsequent responses are fast — typically 2-8 seconds for a 14B model on a modern laptop.

Writing Hindi Prompts — Tips and Examples

OpenClaw works with Hindi prompts. Qwen 2.5 and Llama 3.2 both have reasonable Hindi language support. Here are practical tips:

Tip 1: Use Hinglish for best results. Pure Devanagari Hindi works, but mixing Hindi and English (Hinglish) often produces better outputs because it matches how bilingual Indians actually communicate:

"Mera schedule check karo aur batao aaj kya important meetings hain."

Tip 2: Be explicit about output language. If you want the response in Hindi, say so:

"Ye email ka summary Hindi mein do."

Tip 3: Technical terms in English, instructions in Hindi:

"Is Python code ko review karo aur bugs dhundo. Problems Hindi mein batao."

Example Hindi Prompts:

"Aaj ke liye ek to-do list banao mere calendar ke hisaab se."
(Create a to-do list for today based on my calendar.)

"Ye invoice check karo — GST sahi hai?"
(Check this invoice — is the GST correct?)

"Kal subah 9 baje mujhe reminder do ki Priya ko call karna hai."
(Give me a reminder tomorrow at 9 AM that I need to call Priya.)

"Mere inbox mein jo unread emails hain unka summary do."
(Give me a summary of unread emails in my inbox.)

India-Specific Automations

Here are the OpenClaw automations most relevant to daily life in India:

IRCTC PNR Status Checker

# crons/irctc-check.yml
irctc_pnr_monitor:
  schedule: "0 9 * * *"
  task: "Check PNR status for my upcoming train journeys (read PNR numbers from notes/travel.md) and summarise current booking status"

UPI Transaction Summary

# crons/upi-summary.yml
daily_spend:
  schedule: "0 21 * * *"
  task: "Read today's transactions from finance/upi-log.txt and categorise spending: food, transport, bills, shopping. Write summary to finance/daily-summary.md"

BSE/NSE Market Brief

# crons/market-brief.yml
market_close:
  schedule: "30 15 * * 1-5"    # 3:30 PM IST weekdays (market close)
  task: "Fetch closing prices for my stock watchlist from finance/watchlist.txt and write a brief market summary to finance/market-today.md"

GST Invoice Validator

clawhub install gst-india-validator
openclaw skill enable gst-india-validator
# Usage: openclaw skill run gst-india-validator --file invoice.pdf

Cricket Score Monitor (IPL Season)

cricket_alerts:
  schedule: "*/30 14-23 * * *"    # Every 30 min during match hours
  task: "Check if any IPL match is currently live. If yes, get the current score and send me an update."
  notify: true

Cost Comparison Table

| Setup | Monthly Cost | Privacy | Speed | Best For | |---|---|---|---|---| | Ollama (local) | ₹0 | Complete — data stays on device | Medium (depends on hardware) | Privacy-first, budget users | | Groq Free Tier | ₹0 | Data goes to Groq servers | Very fast (800+ tok/s) | Speed-critical tasks | | Claude API (Sonnet 3.5) | ₹4 per 1K tokens | Data processed by Anthropic | Fast | Complex reasoning | | GPT-4o API | ₹3 per 1K tokens | Data processed by OpenAI | Fast | General complex tasks | | Groq Paid | ₹800/month | Data goes to Groq servers | Fastest | High-volume users |

For most Indian users: start with Ollama (₹0) + Groq free tier (₹0). This covers 95% of personal AI assistant tasks. Add Claude or GPT-4o API credits on a pay-per-use basis only when you need frontier model intelligence for a specific task.

Get Started With OpenClaw Today

You now have everything you need to run a powerful personal AI agent in India for ₹0/month:

1. Install Node.js 22 and OpenClaw
2. Set up Telegram in 5 minutes via BotFather
3. Install Ollama with Qwen 2.5 14B
4. Configure your ABOUT.md with your India context
5. Start with one daily cron — the morning brief is a great first choice

Explore the full OpenClaw tool listing and more India-specific configurations at /tools/openclaw.

For the complete skill ecosystem covering IRCTC, GST, NSE, and more India-specific tools, visit the ClawHub skills directory — search for "india" to filter to locally relevant skills.