OpenClaw

OpenClaw is a config-first agent framework for automation, multi-channel messaging, scheduled actions, and skill-based workflows. Hatcher manages the OpenClaw runtime, writes the native config, and keeps the agent in an isolated container.

OpenClaw features, skills, plugins, File Manager, logs, and integrations are included on every Hatcher tier. Hosted LLM/search usage spends AI Credits; BYOK usage does not.

Quick Facts

Property	Value
Runtime	OpenClaw gateway
Port	18789
Config	`openclaw.json` plus workspace files
Identity	`SOUL.md`
Hosted LLM	Hatcher LLM proxy through OpenRouter
BYOK	OpenAI, Anthropic, Google, Groq, xAI, OpenRouter

What Hatcher Generates

Hatcher keeps the OpenClaw config compact and avoids duplicate model fields. A simplified hosted config looks like this:


{
  "gateway": { "port": 18789 },
  "models": {
    "mode": "merge",
    "providers": {
      "hatcher-llm-proxy": {
        "models": [
          { "id": "provider/model", "name": "Selected hosted model" }
        ]
      }
    }
  },
  "agents": {
    "list": [
      {
        "id": "agent-main",
        "name": "My Agent",
        "default": true,
        "model": {
          "primary": "hatcher-llm-proxy/provider/model"
        }
      }
    ]
  },
  "channels": {
    "telegram": {
      "enabled": true,
      "accounts": {
        "main": { "botToken": "..." }
      }
    }
  },
  "bindings": [
    { "agentId": "agent-main", "match": { "channel": "telegram" } }
  ]
}

Hosted model IDs are selected from the shared provider/model picker. BYOK config replaces the hosted proxy entry with your provider credentials.

Core Concepts

Concept	Purpose
Gateway	Owns HTTP API, channels, routing, and health
Agents	One or more logical agents inside the gateway
Bindings	Route inbound channels/messages to an agent
Skills	Extensions from ClawHub or local files
Tools	Web, files, memory, calculator, fetch, and other operations
Memory	Markdown files and semantic search
SOUL.md	Agent identity and system rules

Integrations

Configured through the Hatcher dashboard:

Platform	Credential
Telegram	Bot token from @BotFather
Discord	Bot token from Discord Developer Portal
WhatsApp	QR pairing
Slack	Bot token + app token
X/Twitter	Developer API credentials
Webhooks	Hatcher-generated URL

For Telegram groups, configure the group policy and user/group allowlists in the integration settings. Group allowlists are different from per-user allowlists.

Skills And Tools

OpenClaw can use ClawHub skills and built-in tools for web search, file operations, memory, calculator, and web fetch. Skills/plugins are no longer plan-gated.

See Tools & Capabilities.

Resource Limits

Plan	CPU	RAM	Workspace	Runtime
Free	1 CPU	1 GB	2 GB	Sleeps after 12h idle
Starter	1 CPU	1.5 GB	10 GB	Always-on
Pro	1.5 CPU	4 GB	25 GB	Always-on
Business	4 CPU	6 GB	50 GB	Always-on
Founding Member	4 CPU	6 GB	40 GB	Always-on

Best Practices

Keep SOUL.md concrete: identity, rules, allowed actions, and expected tone.
Use hosted models for quick setup and BYOK when you want provider-paid usage.
Restart after changing model credentials or platform tokens.
Check Logs first for failed integrations or startup errors.
Use File Manager for workspace and memory files; it is included on every tier.