Everyone's openclaw is building. Yours is struggling.

Answer a few questions. Get a complete config with the right models, cost-saving tips, and capability manifest, so your agent does something.

Try the wizard →

Free to use — pay only when you're ready to download.

Previewing: Windows setup · Also generates for macOS and Linux

Already have OpenClaw installed and running? Skip to Phase 2 to apply your optimized config and install skills.
Part 1 — Open Your Terminal
1
Open PowerShell as Administrator
We need PowerShell with admin privileges to install WSL2. This is the only time you'll use PowerShell — after this, everything runs inside Ubuntu.
How: Click the Start button → Type PowerShell → Right-click → "Run as administrator"
You should see: A blue or black window with white text and a blinking cursor.
2
Install WSL2 (Windows Subsystem for Linux)
This installs a real Linux environment inside Windows. OpenClaw needs Linux to run — it cannot run natively on Windows.
wsl --installCopy
What happens: Windows will download and install Ubuntu. This takes 5-15 minutes. When finished, it will ask you to restart.
3
Create your Linux username and password
The first time Ubuntu opens, it asks you to create a username and password. This is separate from your Windows login.
Part 2 — Install OpenClaw
Part 3 — Add Your API Keys
Part 4 — Apply Your Config & Install Skills
Part 5 — Connect Your Channel
Part 6 — Send Your First Message
// Generated by ClawConfig — Productivity template // Provider: Anthropic · Budget: ~$25/mo · Channel: Telegram { model: "anthropic/claude-sonnet-4-5", modelPreferences: { preferred: "fast", fallbackList: [ "anthropic/claude-haiku-4-5", ], // Aliases: switch mid-chat with /model sonnet, /model haiku models: { "anthropic/claude-sonnet-4-5": { alias: "sonnet" }, "anthropic/claude-haiku-4-5": { alias: "haiku" }, }, }, maxTokens: 40000, memorySearch: { sources: ["memory"], provider: "openai", model: "text-embedding-3-small", // ~$0.10/mo }, channels: { telegram: { enabled: true, botToken: "YOUR_TELEGRAM_BOT_TOKEN", dmPolicy: "pairing", }, }, heartbeat: { interval: 30, model: "anthropic/claude-haiku-4-5", }, // ... cron jobs, tools, compaction settings }
# Agent Capability Manifest
Generated by ClawConfig — edit anytime
## Quick Reference
Provider: Anthropic
Primary: claude-sonnet-4-5 (alias: sonnet)
Fallback: claude-haiku-4-5 (alias: haiku)
Heartbeat: claude-haiku-4-5 every 30m
Budget: ~$25/mo · Daily target: ~$0.83
## Model Assignment
Your config assigns models to roles. Switch manually with /model:
Default (Sonnet): Most tasks — research, writing, analysis, conversation
Haiku: Quick lookups, status checks, simple replies, heartbeat tasks
## Cost Guidelines
Your target budget is approximately $25/month (~$0.83/day)
For expensive tasks, consider switching to Haiku with /model haiku
Always set a hard spending limit on your provider’s dashboard
## Installed Skills
productivity — Calendar, email, task management
web-research — Search and summarize web content
github — Repository monitoring, PR reviews
# Heartbeat Checklist
Generated by ClawConfig — edit anytime, or ask your agent to update this
Check inbox for new emails. Flag anything urgent, draft replies for routine items.
Review calendar for upcoming meetings in the next 4 hours. Prep summaries if needed.
Check GitHub notifications — new issues, PR reviews requested, CI failures.
Scan task list for overdue items. Nudge if anything is more than 24h past due.
Quick memory review — persist anything important from recent conversations.
If nothing needs attention, reply HEARTBEAT_OK.
Security Checklist
API keys are environment variables — never stored in config files. Injected into the systemd service file.
DM policy set to "pairing" — new contacts must be approved with a pairing code before they can message your agent.
Elevated tools restricted — shell commands and file writes require your explicit user ID in the allowlist.
Provider spending limit recommended — set a hard cap on your API provider’s dashboard to prevent runaway costs.
Memory compaction enabled — conversation history is compressed and stored locally, never sent to third parties.
Config files generated client-side — your API keys and configuration never touch our servers.

See the difference

Same agent. Different config.

A default OpenClaw install vs. one configured with ClawConfig. Fifteen seconds tells the whole story.

The problem

OpenClaw is powerful. Setting it up isn't.

Without ClawConfig
  • Install OpenClaw, it asks 19 config questions you don't understand
  • Pick the wrong model, blow through $40 in an afternoon
  • Config is empty defaults — agent can't do anything useful
  • Give up after 2 days
With ClawConfig
  • Answer 6 simple questions — OS, provider, channel, what you want it to do, budget
  • Get an optimized config, capability manifest, and step-by-step guide
  • Agent knows its tools, runs on the right models, and actually works
  • An agent that does what you want it to do

How it works

Three steps. No terminal experience required.

1

Tell us about your setup

Operating system, AI provider, what you want your agent to do, and how much you want to spend. Six screens, two minutes.

2

We generate everything

Optimized config file, capability manifest, heartbeat schedule, and a step-by-step setup guide tailored to your choices.

3

Follow the guide, paste the files

Every terminal command explained. Every click documented. AI helper if you get stuck on any step.

What you get

Four files. One working agent.

openclaw.json

Your complete config

Model assignment, fallback chains, heartbeat schedule, channel setup — all configured for your provider and use case.

clawfy.md

Your agent's brain

Tells your agent what skills it has, which models are available, safety rules, and cost-saving habits.

HEARTBEAT.md

Your agent's daily routine

Inbox checks, overdue tasks, GitHub notifications — customized to the skill packs you selected.

Setup Guide

From zero to running

From "I've never opened a terminal" to "my agent just replied on WhatsApp." Every command, every click, explained.

+ AI helper widget per step
+ Cost estimator
+ 14 days Discord support

BONUS TOOL: Clawfy Pro Extension

Your browser talking to your agent.

Clawfy Extension watches what you're browsing, saves it to your OpenClaw agent's memory, and surfaces the right tools. No more copy paste and explaining what you are researching.

How it works

Install the extension, connect your agent's websocket URL, and browse. Clawfy feeds real-time context to your agent — skill suggestions when you're on GitHub, memory saves when you're researching, conversation awareness across tabs.

Real-time context Skill suggestions Memory saves Conversation awareness
Install Free Extension Upgrade to Pro — $12/mo

Pricing

One price. Everything included.

$49 USD
One-time payment
Try the wizard free →

Just want the extension? $12/mo →

Stuck? Ask a human.

Every purchase includes 14 days of Discord support. Real help from the person who built it — not a bot.

FAQ

Questions

What is OpenClaw?▼
OpenClaw is an open-source AI agent framework with 100k+ GitHub stars. It runs on your machine, connects to messaging platforms like WhatsApp and Telegram, and uses LLMs to handle tasks. ClawConfig generates the configuration files that make it actually useful.
Do I need to know how to code?▼
No. The setup guide walks through every single command — from installing dependencies to starting the agent. The AI helper is there if you get stuck on any step. It's designed for people who have never opened a terminal.
What does it actually cost to run an agent?▼
API costs vary widely — from a few dollars a month with budget models (Gemini Flash, GPT-4o-mini) to $50–100+ with premium models and heavy use. Misconfigured or always-on agents can cost much more. The wizard includes a cost estimator to help you plan, but always set a hard spending limit on your provider’s dashboard.
Can I change my config later?▼
Yes. You can re-run the wizard from your dashboard anytime to regenerate your config with different choices, or edit the files directly if you're comfortable.
What if I get stuck?▼
Every step of the setup guide has an AI helper widget you can ask questions. You also get 14 days of Discord support where you can get help directly.
What's the refund policy?▼
Refund policy details coming soon. Contact us if you have any issues with your purchase.

Ready to make your agent useful?

Try the wizard for free. Pay when you're happy with the config.

Try the wizard →