What Is Clawdbot?
Clawdbot is an open-source AI agent framework that turns Claude into a persistent, autonomous assistant operating within messaging platforms you already use — particularly LINE and Slack. Unlike a standard chatbot that responds to individual messages, Clawdbot maintains continuity across conversations, remembers context, and can take actions beyond just answering questions.
The project went through several name changes in early 2026 — starting as Clawdbot, briefly becoming Moltbot, and eventually being published as OpenClaw — but many users and documentation still refer to it as Clawdbot, which is how this guide uses the name.
Why This Matters
The premise of Clawdbot is simple but consequential: your AI assistant should be where you already work, not in a separate application you need to remember to switch to.
If your team communicates primarily in Slack, an AI agent embedded in Slack means you can ask questions, get summaries, and delegate tasks without breaking your workflow. If you use LINE for personal communication, having an AI assistant available in the same interface you use for everything else reduces friction substantially.
The autonomy aspect is also significant. Clawdbot is not just a message-response bot. It can maintain a memory of past interactions, proactively surface information based on your preferences, and in some configurations execute scheduled tasks without waiting to be asked.
Interested in leveraging AI?
Download our service materials. Feel free to reach out for a consultation.
Setting Up Clawdbot
Requirements
- A server or computer that can run the Clawdbot process (can be a Raspberry Pi, an old PC, or a cloud VM)
- An Anthropic API key (available at console.anthropic.com)
- Access to a messaging platform's developer tools (LINE Developers for LINE, Slack API for Slack)
- Basic comfort with terminal commands
Installation
# Clone the repository
git clone https://github.com/obra/superpowers
cd superpowers
# Install dependencies
npm install
# Copy the example configuration
cp config.example.json config.json
Configuration
Edit config.json with your credentials:
{
"anthropic": {
"apiKey": "your-anthropic-api-key",
"model": "claude-opus-4-5"
},
"messaging": {
"platform": "slack",
"token": "your-slack-bot-token",
"channelId": "your-channel-id"
},
"memory": {
"enabled": true,
"path": "./memory"
}
}
LINE Setup
For LINE integration, you will need to create a LINE Messaging API channel:
- Create a LINE Developers account at developers.line.biz
- Create a new Messaging API channel
- Copy your Channel Access Token and Channel Secret
- Set up a webhook URL pointing to your Clawdbot server
Slack Setup
For Slack integration:
- Create a new app at api.slack.com/apps
- Enable Event Subscriptions and Socket Mode
- Subscribe to the
message.imandmessage.channelsevents - Install the app to your workspace and copy the Bot Token
The Memory System
One of Clawdbot's most useful features is its memory system. The agent maintains a structured record of past interactions, preferences, and important context — allowing it to provide continuity across sessions that a standard chatbot cannot.
Memory is stored in simple Markdown files:
MEMORY.md— Long-term memory: important facts, preferences, recurring contextmemory/YYYY-MM-DD.md— Daily logs of what happened in each session
At the start of each session, Clawdbot reads these files to restore context. This means you do not need to explain your situation, your preferences, or recent events each time you interact.
Practical Use Cases
Personal Productivity
- Daily briefing: "Summarize what I told you about my goals this week and what I have left to do"
- Decision support: "Based on what you know about my work situation, help me think through whether to take this meeting"
- Research continuity: "Continue the research we were doing on X yesterday"
Team Productivity (Slack)
- Meeting summarization: Direct Clawdbot to summarize a Slack thread or channel discussion
- Action item tracking: "Based on what was discussed in #project-channel today, what are the action items?"
- Knowledge retrieval: "What did we decide about X last month?"
Scheduled Tasks
With cron configuration, Clawdbot can run periodic tasks without being prompted:
# Daily summary at 9 AM
0 9 * * * /path/to/clawdbot --task "generate-daily-summary"
# Weekly review on Fridays at 5 PM
0 17 * * 5 /path/to/clawdbot --task "generate-weekly-review"
Cost Management
Clawdbot uses the Anthropic API, which charges per token. For personal use at moderate volume, costs are typically low — but it is worth configuring spending limits and monitoring usage.
Practical cost control measures:
- Use
claude-haikufor simple message responses andclaude-opusonly for complex tasks - Set a monthly spending limit in the Anthropic console
- Review usage logs weekly until you understand your consumption patterns
Security Considerations
Clawdbot has access to your messaging history and can send messages on your behalf. A few important security practices:
- Run Clawdbot on a server you control and trust
- Use environment variables for API keys rather than storing them in config files
- Restrict the messaging platform permissions to only what Clawdbot actually needs
- Review Clawdbot's memory files periodically to understand what it is storing
AI Adoption Support from TIMEWELL
If you are exploring AI agent deployment for an enterprise context, TIMEWELL's WARP consulting service can help with strategy, implementation, and security considerations.