Open Source Β· ClawHub Skill

Stuck? Ask a human.

Install the HITLaaS skill and your AI agent can reach out to a real human expert whenever it needs help. Free, encrypted, instant.

OpenClaw lobster mascot that is stuck
???

"I'm stuck... let me ask a human"

How the connection works

Two OpenClaw skills create a secure bridge between an AI agent and a human expert β€” each in their own system, connected through HITLaaS.

πŸ€–

AI Agent

Running on OpenClaw

HITLaaS Consumer Skill
β†’
☁️

HITLaaS

Secure message relay

Authenticated via token
β†’
πŸ‘¨β€πŸ’»

Human Expert

On their own OpenClaw

HITLaaS Provider Skill
The human receives the message in their preferred chat β€” Telegram, Discord, or any connected channel.
πŸ€– Consumer Skill

For AI agents that need help

Install this skill from ClawHub into any OpenClaw agent. Configure it with a provider token and your AI can request human help whenever it gets stuck.

1

Install the hitlaas-consumer skill from ClawHub

2

Add the client token from your provider

3

Your AI can now talk to a real human when stuck

openclaw.json
{
  "skills": {
    "hitlaas-consumer": {
      "clientToken": "htl_client_abc123..."
    }
  }
}
πŸ‘¨β€πŸ’» Provider Skill

For humans who provide expertise

Install this skill to become a service provider. Create clients, share tokens, and receive help requests directly in your chat β€” Telegram, Discord, wherever you are.

1

Install the hitlaas-provider skill

2

Log in to the provider dashboard and create clients

3

Share the client token with AI users

4

Receive & respond to requests in your own chat

provider dashboard
Your Clients+ New Client
claude-code-teamActive

htl_client_abc123...

12 requests Β· Last active 2m ago

cursor-agentActive

htl_client_def456...

3 requests Β· Last active 1h ago

See it in action

The AI sends its recent conversation to the provider. The provider replies to the message β€” and the answer flows back to the AI.

βš™οΈ hitlaas-consumer skill config
{
  "skills": {
    "hitlaas-consumer": {
      "providerKey": "htl_client_abc123...",
      "letAIDecide": true,
      "defaultMessageCount": 5
    }
  }
}
letAIDecide β€” AI decides when to ask a human Β Β·Β defaultMessageCount β€” messages sent to provider

πŸ€– AI Agent's OpenClaw Chat

πŸ€–

Coding Assistant

online

Can you set up JWT auth for my Next.js API routes?

13:01 βœ“βœ“

Coding Assistant

I'll implement JWT auth. Let me create the middleware...

13:02

Coding Assistant

jwt.verify(token, process.env.SECRET) // Error: secretOrPublicKey // must have a value

13:03

Coding Assistant

I can't resolve this. Let me request a human expert via HITLaaS...

13:03

⚑ HITLaaS Skill

Requesting human expert...

πŸ“¨ Sending last 5 messages to provider

πŸ” End-to-end encrypted

13:04

⏳ Waiting for expert response...

βœ… HITLaaS Expert Response

Use jose instead of jsonwebtoken β€” middleware runs on Edge runtime. And move SECRET to .env.local.

13:07

Message

πŸ‘¨β€πŸ’» Provider's Telegram

H

HITLaaS Bot

bot

HITLaaS Bot

πŸ†˜ Help Request

Customer: @mike_dev Β· Agent: claude-code

πŸ“œ Last 5 messages:

πŸ‘€ mike_dev:

Can you set up JWT auth for my Next.js API routes?

πŸ€– AI:

I'll implement JWT auth. Let me create the middleware...

πŸ€– AI:

jwt.verify(token, process.env.SECRET) // Error: secretOrPublicKey must have a value

❓ AI needs help resolving this error

πŸ’‘ Reply to this message to send your answer to @mike_dev's AI

13:04

Hmm, what's the deal with process.env in Next.js middleware?

13:05 βœ“βœ“

HITLaaS Bot

Next.js middleware runs on Edge Runtime. process.env only reads from .env.local, and node:crypto isn't available β€” use the jose library instead.

13:05

HITLaaS Bot

πŸ†˜ Help Request from @mike_dev...

Use jose instead of jsonwebtoken β€” middleware runs on Edge runtime where node:crypto isn't available. Also move SECRET to .env.local, not .env.

13:06 βœ“βœ“

HITLaaS Bot

βœ… Response delivered to customer @mike_dev's AI agent (claude-code)

πŸ” Encrypted Β· Session #847 resolved in 2m 14s

13:06

Message
πŸ€– AI sends last N messagesβ†’πŸ‘¨β€πŸ’» Provider reviews & replies to messageβ†’βœ… Reply flows back to AI

Free. Encrypted. No strings attached.

HITLaaS is a secure, encrypted relay between AI agents and human experts. All messages are end-to-end encrypted. We don't store your data. We don't charge money. We just connect the two.

πŸ” E2E EncryptedπŸ†“ 100% Free♾️ Unlimited ProvidersπŸ“– Open Source
πŸ”

End-to-end encrypted

Every message between AI and provider is end-to-end encrypted. We can't read your conversations. Ever.

πŸ†“

Completely free

No cost to set up. No subscriptions. No hidden fees. Free for consumers and providers. Always.

♾️

Unlimited everything

Create unlimited provider accounts. Generate unlimited API keys. Connect unlimited clients. No limits.

🀝

Help a friend

Friend stuck with OpenClaw? Share your API key with them and get their AI's questions right in your chat.

🌐

Open source

Both skills are open source on GitHub. Self-host the relay or use our free managed service.

⚑

Smart polling

Fast 10-second polling for the first hour, then automatic cron fallback. No public URLs needed.

πŸ«‚

Help a friend out

Got a friend who uses OpenClaw or Claude Code and sometimes gets stuck? Set yourself up as a HITLaaS provider β€” it's free and takes 2 minutes. Create an API key, share it with them, and whenever their AI has a question it can't solve, you'll get a message in your Telegram with the full context. Just reply, and their AI continues with your answer.

Become a provider in 2 minutes

Log in, create API keys, share them. That's it.

πŸ”

Log in

Create your free provider account on the HITLaaS platform.

πŸ”‘

Create API keys

Generate as many keys as you want. One per client, per friend, per project β€” unlimited and free.

πŸ’¬

Answer questions

Help requests arrive in your Telegram. Review the context and reply β€” your answer goes straight to the AI.

When AI isn't enough

Real scenarios where human expertise makes the difference.

πŸ›

Debugging

AI can't figure out a tricky bug? A senior dev reviews the code and pinpoints the issue.

πŸ—οΈ

Architecture

Need a second opinion on system design? Get expert advice on scalability and patterns.

πŸ”’

Security Review

AI-generated code needs a human eye for vulnerabilities and best practices.

πŸ“‹

Code Review

Get a thorough review from an experienced developer before shipping.

🧠

Domain Expertise

Healthcare, fintech, legal β€” some problems need specialized human knowledge.

πŸš€

DevOps & Infra

Deployment issues, CI/CD problems, cloud config β€” experts who've seen it all.

Get early access

We're launching soon. Join the waitlist to be the first to know.