Slack
AI assistance where your team already works
The Fastest Way: HeraClaw Cloud ā”
Skip the Slack app setup, OAuth configuration, and bot token management. HeraClaw Cloud comes with Slack integration pre-configured and ready to use in 60 seconds.
Why HeraClaw Cloud for Slack?
ā Ready in 60 seconds - No Slack app creation, no OAuth scopes, no bot token management ā Pre-configured bot - Works immediately after workspace authorization ā Always updated - We handle Slack API changes and deprecations automatically ā Professional support - If anything breaks, we fix it for you ā Secure by default - Enterprise-grade security, SOC 2 compliant infrastructure ā Zero maintenance - No server management, no SSL certificates, no uptime monitoring
How it works:
- Sign up at cloud.getopenclaw.ai (takes 60 seconds)
- Go to Integrations ā Slack
- Click 'Connect to Slack'
- Authorize our bot for your workspace
- Done! Start using your AI assistant in Slack
Get Started: Start with HeraClaw Cloud ā
Complete Guide to OpenClaw + Slack
Slack is where enterprise teams live. With 10+ million daily active users across organizations worldwide, it's the central nervous system of modern work. OpenClaw's Slack integration brings powerful AI assistance directly into this workflow hub, eliminating context-switching and making AI help available exactly where teams collaborate.
Whether you're running a 10-person startup or a 10,000-employee enterprise, this integration transforms how your team accesses AI assistance. No new apps to learn, no separate interfaces - just @mention your assistant in any channel and get instant, intelligent help.
Why Use Slack with OpenClaw?
Slack's architecture makes it one of the most powerful platforms for AI assistants. Here's why the OpenClaw + Slack combination is transformative:
1. Enterprise-Grade Permission Control
Slack's permission system was built for enterprise security, and OpenClaw leverages it completely:
- Channel-level control: Restrict the assistant to specific channels like #ai-help or #engineering
- Role-based access: Only users with certain roles (Team, Engineering, Leadership) can interact
- Private channel support: Works in private channels with full context awareness
- DM privacy: One-on-one conversations for sensitive questions
- Workspace isolation: Different configs per Slack workspace for multi-tenant orgs
- Audit logs: Track every interaction for compliance and security review
This makes it perfect for enterprises where security and compliance are critical. Your legal team can review exactly who asked what, when.
2. @Mention Interface - Natural and Familiar
No slash commands to remember, no special syntax. Just @mention your assistant like you'd @mention a colleague:
@OpenClaw what's on my calendar today?@OpenClaw summarize the last 20 messages in #product-updates@OpenClaw create a task to follow up with the client by FridayYour team already knows how to use it. Zero training required.
3. Thread-Based Conversations
Slack's thread system pairs perfectly with AI conversations:
- Deep dives in threads: Start a thread from any message for extended conversations
- Channel cleanliness: Main channel stays clean while detailed help happens in threads
- Context preservation: Assistant remembers everything discussed in the thread
- Multiple concurrent conversations: Different team members can have parallel discussions without crosstalk
- Searchable history: All AI interactions are searchable in Slack's search
Example: Someone asks a complex technical question in #engineering. The assistant starts a thread with the answer, and multiple engineers can ask follow-up questions in that same thread. The main channel just shows one message.
4. Rich Message Formatting for Technical Content
Slack's message formatting makes technical information readable:
- Code blocks with syntax highlighting: Perfect for engineering teams
- Formatted lists: Bullet points and numbered lists for clarity
- Inline code:
variable namesandfunction()calls are properly formatted - Links and references: Clickable links to docs, PRs, and resources
- Blockquotes: Highlight important information
- Emoji reactions: Quick feedback without cluttering threads
Technical answers look professional and are easy to scan. No more walls of unformatted text.
5. File Upload and Analysis
Upload files directly to Slack for AI analysis:
- Images: "Explain this architecture diagram" "What's wrong in this screenshot?"
- Code files: "Review this Python script" "Find security issues in this code"
- Documents: "Summarize this contract" "Extract action items from this PDF"
- Spreadsheets: "What's the total revenue in Q4?"
- Logs: "Parse this error log and explain what failed"
File size limits: Up to 1GB per file on paid Slack plans. The assistant can read and analyze everything.
6. Slash Commands for Structured Operations
For power users, custom slash commands provide quick access to common operations:
/ask [question]- Direct question without @mention/brief- Get your personalized daily briefing/task [description]- Create a task in your system/search [query]- Search across your knowledge base/summarize- Summarize recent channel messages/code [language] [description]- Generate code snippets
Slash commands feel native to Slack and provide auto-completion as you type.
7. Multi-Workspace Support for Enterprise
Enterprise organizations often have multiple Slack workspaces (one per business unit, region, or team). OpenClaw supports this:
- Same bot across workspaces: One OpenClaw instance, multiple Slack workspaces
- Different permissions per workspace: Engineering workspace gets code access, sales workspace gets CRM access
- Shared or isolated memory: Choose whether workspaces share context or stay isolated
- Centralized management: Manage all workspace integrations from one dashboard
Perfect for companies that operate multiple Slack workspaces but want consistent AI assistance everywhere.
8. Enterprise Grid Support
For organizations using Slack Enterprise Grid:
- Org-wide deployment: Deploy across all workspaces in your Grid
- Centralized app management: IT admins approve once, available everywhere
- Cross-workspace search: Assistant can search across workspaces (if configured)
- Compliance and eDiscovery: All AI interactions are captured in Grid's compliance exports
Real-World Use Cases
1. Engineering Team Support
Scenario: Your engineering team uses Slack for all communication. Developers constantly ask questions about your codebase, deployment procedures, and debugging approaches. Senior engineers spend 30-40% of their time answering repetitive questions.
How OpenClaw helps:
Add your assistant to #engineering and #help channels. When developers ask:
- "How do I run tests locally?" ā Assistant pulls from your docs and responds with exact commands
- "What's the deploy process for staging?" ā Step-by-step instructions with links to runbooks
- "Why is this API endpoint returning 403?" ā Assistant checks recent deployments, reviews auth logic, suggests likely causes
- "Show me examples of pagination in our codebase" ā Finds and shares relevant code snippets
Impact:
- Senior engineers reclaim 10-15 hours per week
- New hires get instant answers instead of waiting for responses
- Documentation stays up-to-date because the assistant can flag outdated info
- Context-switching reduced - answers appear in Slack instead of requiring Confluence/Notion
2. Sales Team Enablement
Scenario: Your sales team lives in Slack. They need quick access to product info, pricing details, competitive intelligence, and customer context during calls and prospect conversations.
How OpenClaw helps:
Integrate with your CRM, knowledge base, and product docs:
- During discovery calls: "@OpenClaw what features do we have for enterprise SSO?" ā Instant product capability summary
- Pricing questions: "What's our pricing for 500-seat deals?" ā Pulls from pricing database with approval workflows
- Competitive intel: "How do we compare to Competitor X on feature Y?" ā Battle card summary
- Customer context: "Give me the history with Acme Corp" ā Recent emails, deal status, past conversations
- Proposal generation: "Draft a proposal for Acme Corp, 200 seats, annual billing" ā Creates first draft
Impact:
- Sales reps close deals 20% faster with instant information access
- Fewer lost deals due to delayed responses during prospect conversations
- Consistent messaging across the sales team
- CRM stays updated without manual data entry
3. Customer Support Automation
Scenario: Your customer support team uses Slack for internal coordination. They need to quickly find answers to customer questions, check system status, and escalate issues.
How OpenClaw helps:
Connect to your knowledge base, ticketing system, and monitoring tools:
- Knowledge base queries: "How do customers reset their password?" ā Pulls from help docs
- System status: "Is the API down?" ā Checks monitoring tools and recent incidents
- Customer history: "What tickets has customer@company.com filed?" ā Recent ticket summary
- Escalation help: "Create a P1 incident for database outage" ā Creates ticket and notifies on-call
- Response drafting: "Draft a response explaining the outage" ā Professional, empathetic response template
Impact:
- First response time drops from 30 minutes to under 2 minutes
- Support agents handle 50% more tickets with the same team size
- Customer satisfaction increases due to faster, more accurate responses
- Knowledge base gaps identified automatically (assistant flags missing docs)
4. Product Team Coordination
Scenario: Your product team coordinates feature development, user research, and roadmap planning in Slack. Information is scattered across Notion, Jira, user interviews, and Slack threads.
How OpenClaw helps:
- Roadmap queries: "What features are planned for Q2?" ā Pulls from Jira roadmap
- User research synthesis: "Summarize feedback about the new onboarding flow" ā Analyzes recent user interviews and support tickets
- Feature status: "Where are we on the mobile redesign?" ā Jira query + recent Slack discussion summary
- Decision logging: "Document our decision to use React over Vue" ā Creates decision record in Notion
- Meeting prep: "Summarize discussion in #product-ideas about AI features" ā Thread summaries with key points
Impact:
- Product decisions made 40% faster with instant access to all context
- No more "where did we discuss that?" lost information
- Automatic documentation of key decisions
- User feedback surfaces more consistently in planning
5. Executive Daily Briefings
Scenario: Your leadership team needs high-level visibility without drowning in Slack messages. They want to know what's important without reading every channel.
How OpenClaw helps:
Configure automated morning briefings delivered via DM:
Good morning! Here's your Friday briefing:š„ Urgent (needs your attention):- Server outage in #incidents (resolved, postmortem needed)- Enterprise deal with Acme Corp needs approval (see #sales-urgent)š Metrics:- MRR: $847K (+3.2% vs last week)- Active incidents: 0- Team velocity: 32 points this sprint (on track)š¬ Key discussions:- #engineering: Debating migration to Kubernetes (12 messages)- #product: Feature prioritization for Q2 (thread with 18 replies)- #sales: 3 deals closing this week, $140K total ARRš
Your calendar:- 10am: Board prep with CFO- 2pm: All-hands presentation- 4pm: 1:1 with Head of EngineeringAnything you want to dive deeper on?Impact:
- Executives stay informed without reading 500+ Slack messages per day
- Critical issues surface immediately
- Context available for all meetings
- Team feels heard (execs are aware of key discussions)
6. HR and People Operations
Scenario: Your HR team answers repetitive questions about benefits, policies, PTO, and company procedures. New employees constantly ask the same questions.
How OpenClaw helps:
Add the assistant to #hr-help or #ask-people-ops:
- "How much PTO do I have left?" ā Checks HR system and responds
- "What's our parental leave policy?" ā Pulls from employee handbook
- "How do I submit an expense report?" ā Step-by-step guide with links
- "Who do I contact about health insurance?" ā Contact info and hours
- "What holidays do we observe?" ā Current year holiday calendar
Impact:
- HR team spends 70% less time answering repetitive questions
- Employees get answers 24/7, not just during business hours
- Policy information always up-to-date (single source of truth)
- New hires onboard faster with instant access to information
7. DevOps and Incident Response
Scenario: Your DevOps team manages incidents in Slack. During outages, every second counts. Information needs to flow fast.
How OpenClaw helps:
Integrate with monitoring, logging, and deployment tools:
- Incident detection: "Alert: API latency above threshold" ā Assistant creates incident channel and notifies on-call
- Status queries: "What's the current error rate?" ā Queries Datadog/New Relic and reports
- Log analysis: "Show me errors in the last 10 minutes" ā Fetches and summarizes logs
- Deployment info: "What was deployed in the last hour?" ā Lists recent deployments
- Runbook access: "How do we roll back a deployment?" ā Links to runbook with commands
- Postmortem creation: "Create a postmortem for today's incident" ā Generates template with timeline
Impact:
- Mean time to resolution (MTTR) decreases 30-40%
- Fewer escalations needed (junior engineers can resolve more issues)
- Incident documentation happens automatically
- Knowledge captured and searchable for future incidents
8. Cross-Functional Project Management
Scenario: You have a major project spanning Engineering, Product, Design, and Marketing. Coordination happens across multiple Slack channels and tools.
How OpenClaw helps:
Create a project-specific assistant instance:
- Status updates: "What's the status of the website redesign?" ā Checks Jira, GitHub, and Figma for progress
- Blocker identification: "What's blocking the launch?" ā Scans project channels and tickets for blockers
- Cross-team coordination: "Has marketing approved the new messaging?" ā Checks approval thread in #marketing
- Timeline queries: "Are we on track for the Q2 launch?" ā Compares progress to timeline
- Stakeholder updates: "Generate a status update for leadership" ā Synthesizes progress across all workstreams
Impact:
- Project visibility increases across all stakeholders
- Fewer status update meetings needed (async via assistant)
- Blockers surface faster
- Launch timelines more accurate with real-time tracking
Features Deep Dive
@Mention Support
Mention your assistant anywhere in Slack:
- In channels it's invited to
- In threads for context-aware help
- In DMs for private assistance
- Multiple mentions per message supported
- Works with display name and handle
Slash Commands
Create custom slash commands for your team:
/ask- Quick questions without @mention/task- Create tasks in your system/search- Search knowledge base/brief- Daily/weekly briefings/status- Check project/system status- Custom commands you define
Channel Permission Control
Fine-grained control over where the assistant operates:
- Allowlist specific channels (e.g., only #ai-help, #engineering)
- Blocklist channels where it should never respond (e.g., #leadership, #hr-private)
- Private channel support with opt-in
- DM support toggle (enable/disable DMs)
- Per-channel different capabilities (code access in #engineering, CRM access in #sales)
Thread Intelligence
Full thread context awareness:
- Remembers entire thread history
- Responds in threads to keep channels clean
- Multiple people can participate in thread conversations
- Thread summaries on demand
- Export thread conversations to docs
File Processing
Upload and analyze files:
- Images: Screenshots, diagrams, charts, photos
- Code: Python, JavaScript, Java, Go, any language
- Documents: PDFs, Word, Google Docs exports
- Spreadsheets: CSV, Excel, Google Sheets exports
- Logs: Application logs, server logs, error reports
- Archives: ZIP files (extracts and analyzes contents)
Size limits: Up to 1GB per file on paid Slack plans
Rich Formatting
Professional message formatting:
- Code blocks with syntax highlighting
- Inline code formatting
- Bulleted and numbered lists
- Blockquotes for emphasis
- Links (text and buttons)
- Emoji for visual clarity
- Tables (via code blocks)
- Message blocks for structured info
Multi-Workspace
Manage multiple Slack workspaces:
- One assistant instance, many workspaces
- Different permissions per workspace
- Shared or isolated memory
- Centralized dashboard for management
- Workspace-specific configurations
DM Support
Private one-on-one conversations:
- Full feature access via DMs
- Personal context and history
- Secure for sensitive questions
- Works across all devices
- No channel permissions apply
Enterprise Grid
For large organizations:
- Org-wide deployment
- Centralized app approval
- Cross-workspace functionality
- Compliance export integration
- eDiscovery support
Security Features
- SOC 2 Type II certified (HeraClaw Cloud)
- Encryption in transit (TLS 1.3)
- Encryption at rest (AES-256)
- OAuth 2.0 authentication
- Role-based access control
- Audit logging (all interactions)
- Data retention controls
- GDPR compliant
Setup Option 1: HeraClaw Cloud (Recommended)
Time required: 60 seconds Technical skill: None Cost: Included in HeraClaw Cloud subscription Best for: 95% of users, all team sizes
Why HeraClaw Cloud?
- No Slack app configuration required
- No OAuth scope management
- No bot token security concerns
- No server infrastructure to maintain
- No SSL certificates to renew
- No uptime monitoring needed
- Professional support included
- Automatic updates and security patches
- 99.9% uptime SLA
- Enterprise-grade security
Steps:
1. Sign up for HeraClaw Cloud
- Visit cloud.getopenclaw.ai
- Click "Sign In" (no credit card required)
- Create your account (takes 60 seconds)
2. Navigate to Integrations
- Click "Integrations" in the left sidebar
- Find "Slack" in the list
- Click "Connect to Slack"
3. Authorize the Workspace
- Slack OAuth screen appears
- Select your workspace from dropdown
- Review permissions (you can customize)
- Click "Allow"
4. Configure Channels (Optional)
- Back in HeraClaw dashboard
- Choose which channels the assistant can access
- Set permissions (who can interact)
- Save configuration
5. Invite to Channels
- In Slack, go to any channel
- Type
/invite @OpenClaw - The bot appears in the channel
6. Test It
- Type
@OpenClaw hello! - You should get an immediate response
- Try asking a question:
@OpenClaw what can you do?
That's it! You're up and running.
What You Get with HeraClaw Cloud:
ā Instant Setup - No technical configuration ā Automatic Updates - We handle Slack API changes ā Professional Support - Email, chat, and phone support ā 99.9% Uptime - SLA-backed reliability ā Enterprise Security - SOC 2, GDPR, HIPAA available ā Unlimited Workspaces - Connect as many as you need ā Advanced Features - Multi-workspace, custom slash commands, audit logs ā No Maintenance - We manage servers, scaling, monitoring
Pricing: See cloud.getopenclaw.ai/pricing (starts with affordable team plans)
Get Started: Start with HeraClaw Cloud ā
Setup Option 2: Self-Hosted (Advanced)
Time required: 30-45 minutes (first time), 15-20 minutes (if experienced) Technical skill: Intermediate to Advanced Cost: VPS hosting ($10-50/month) + your time Best for: DevOps engineers, technical teams, compliance requirements
Who should self-host?
ā DevOps engineers who enjoy infrastructure ā Organizations with strict data residency requirements ā Teams already running Kubernetes/Docker infrastructure ā Technical enthusiasts with homelab setups ā Companies that cannot use third-party SaaS
Who should NOT self-host?
ā Small teams without DevOps expertise ā Anyone who values time over cost savings ā Teams without 24/7 on-call coverage ā Organizations without security/compliance teams
Prerequisites:
- OpenClaw installed and running (Mac/Linux/VPS)
- Terminal/SSH access to your OpenClaw server
- Slack workspace admin permissions
- Understanding of OAuth 2.0
- Basic networking knowledge (ports, firewalls, reverse proxies)
- SSL certificate (for production)
Detailed Self-Hosted Setup
Step 1: Create a Slack App
- Go to api.slack.com/apps
- Click "Create New App"
- Choose "From scratch"
- App Name: "OpenClaw Assistant" (or your preference)
- Pick your workspace
- Click "Create App"
Step 2: Enable Socket Mode (Recommended for Most Users)
Socket Mode avoids needing a public URL with SSL.
- In your Slack app settings, go to "Socket Mode"
- Toggle "Enable Socket Mode" to ON
- Click "Generate an app-level token"
- Token Name: "OpenClaw Socket"
- Add scope:
connections:write - Click "Generate"
- Copy this token (starts with
xapp-) - you'll need it later
With Socket Mode, you don't need to expose OpenClaw to the internet. Recommended for self-hosted setups.
Step 3: Configure OAuth Scopes
- Go to "OAuth & Permissions" in sidebar
- Scroll to "Scopes" section
- Under "Bot Token Scopes", add:
Required scopes:
app_mentions:read- See when bot is @mentionedchat:write- Send messageschannels:history- Read channel messageschannels:read- View channel infogroups:history- Read private channel messages (if needed)groups:read- View private channelsim:history- Read DM messagesim:read- View DMsim:write- Send DMsusers:read- Get user information
Optional scopes (for advanced features):
files:read- Read uploaded filesfiles:write- Upload filesreactions:read- Read emoji reactionsreactions:write- Add emoji reactionscommands- Enable slash commands
- Click "Save Changes"
Step 4: Subscribe to Events
-
Go to "Event Subscriptions"
-
Toggle "Enable Events" to ON
-
If using Socket Mode: You're done with this step
-
If NOT using Socket Mode (advanced):
- You need a public URL with SSL
- Enter Request URL:
https://your-openclaw-domain.com/slack/events - Slack will verify the URL (OpenClaw must be running)
-
Scroll to "Subscribe to bot events"
-
Add these events:
app_mention- When someone @mentions your botmessage.channels- Messages in public channelsmessage.groups- Messages in private channelsmessage.im- Direct messages to bot
- Click "Save Changes"
Step 5: Install App to Workspace
- Go to "Install App" in sidebar
- Click "Install to Workspace"
- Review permissions
- Click "Allow"
- Copy the "Bot User OAuth Token" (starts with
xoxb-) - Keep this secure - it's the key to your Slack workspace
Step 6: Create Slash Commands (Optional)
- Go to "Slash Commands" in sidebar
- Click "Create New Command"
- Example command:
- Command:
/ask - Request URL: Your endpoint (handled by Socket Mode or your URL)
- Short Description: "Ask the AI assistant anything"
- Usage Hint: "[your question]"
- Command:
- Click "Save"
- Repeat for other commands (
/brief,/task,/search, etc.)
Step 7: Configure OpenClaw
Edit your OpenClaw config file (usually ~/.openclaw/config.yaml or ~/.config/openclaw/config.yaml):
channels: slack: enabled: true # Bot token from Step 5 botToken: "xoxb-your-bot-token-here" # App-level token from Step 2 (if using Socket Mode) appToken: "xapp-your-app-token-here" # Use Socket Mode (recommended) socketMode: true # Optional: Restrict to specific channels allowedChannels: - "C01234567" # #ai-help - "C09876543" # #engineering # Optional: Never respond in these channels ignoredChannels: - "C11111111" # #leadership - "C22222222" # #random # Optional: Only these users can interact (user IDs) allowedUsers: - "U0123456789" # Optional: Respond to @mentions (default: true) appMention: true # Optional: Respond in DMs (default: true) directMessage: true # Optional: Respond in threads (default: true) threads: true # Optional: Slash command prefix commandPrefix: "/"Finding Channel and User IDs:
- In Slack, enable Developer Mode:
- Slack ā Preferences ā Advanced ā "Show Developer mode" ā
- Right-click any channel ā "Copy Link"
- The URL contains the channel ID:
/archives/C01234567 - Right-click any user ā "Copy member ID"
Step 8: Start OpenClaw
# If running as a servicesudo systemctl restart openclaw# If running directlyopenclaw gateway start# Check logsopenclaw gateway logs --followWhat to look for in logs:
ā Slack Socket Mode connectedā Bot user: @OpenClaw (ID: B01234567)ā Listening for mentions, DMs, and commandsā Serving 1 workspace(s)Step 9: Invite Bot to Channels
In Slack:
- Go to any channel you want the bot in
- Type
/invite @OpenClaw - Bot appears in channel member list
Repeat for all desired channels.
Step 10: Test the Integration
-
In a channel where you invited the bot:
bash@OpenClaw hello! -
You should get a response within 2 seconds
-
Try a DM:
- Open DM with your bot
- Send: "What can you do?"
- Should respond with capabilities
-
Test a slash command (if configured):
bash/ask What's the weather like today?
If it doesn't respond, check logs for errors.
Advanced Self-Hosted Configuration
Multiple Workspaces
Run one OpenClaw instance across multiple Slack workspaces:
channels: slack: instances: main_workspace: botToken: "xoxb-token-workspace-1" appToken: "xapp-token-workspace-1" socketMode: true allowedChannels: - "C01234567" client_workspace: botToken: "xoxb-token-workspace-2" appToken: "xapp-token-workspace-2" socketMode: true # Different config for this workspaceCustom Slash Commands
Map slash commands to specific OpenClaw actions:
channels: slack: commands: ask: description: "Ask the AI anything" handler: "chat" brief: description: "Get your daily briefing" handler: "briefing" task: description: "Create a task" handler: "task_create" options: - name: "description" required: true - name: "due_date" required: falseChannel-Specific Configurations
Different behavior per channel:
channels: slack: channelConfigs: # Engineering channel: code access enabled "C01234567": allowedSkills: - "code" - "github" - "terminal" modelPreference: "claude-opus-4" # Sales channel: CRM access "C09876543": allowedSkills: - "crm" - "email" modelPreference: "claude-sonnet-4"Rate Limiting
Protect your OpenClaw instance from abuse:
channels: slack: rateLimit: # Per user limits perUser: requests: 20 window: 60 # seconds # Per channel limits perChannel: requests: 100 window: 60 # Global limits global: requests: 1000 window: 60Response Formatting
Control how messages appear:
channels: slack: formatting: # Max message length before splitting maxLength: 3000 # Split long messages? splitLongMessages: true # Use threads for follow-ups? threadReplies: true # Include thinking indicators? showTyping: true # Code block language detection detectCodeLanguage: true # Add emoji to responses? emoji: trueTroubleshooting Self-Hosted Setup
Bot Not Responding to @Mentions
Symptoms: You @mention the bot, nothing happens
Potential causes:
- Event Subscriptions not configured - Check "Event Subscriptions" in Slack app settings, ensure
app_mentionevent is subscribed - Bot not in channel - Type
/invite @YourBotin the channel - OpenClaw not running - Check
openclaw gateway status - Socket Mode connection failed - Check logs for connection errors
- Wrong bot token - Verify
botTokenin config matches Slack
Fix:
# Check OpenClaw is runningopenclaw gateway status# Check logs for errorsopenclaw gateway logs --tail 50# Restart OpenClawopenclaw gateway restartSlash Commands Not Appearing
Symptoms: Type / in Slack, your commands don't show up
Potential causes:
- Commands not created in Slack app settings
- Commands need to be reinstalled
- Slack is caching old command list
- Wrong workspace
Fix:
- Go to api.slack.com/apps ā Your App ā Slash Commands
- Verify commands exist
- Reinstall app: "Install App" ā "Reinstall to Workspace"
- Wait 5 minutes for Slack to refresh cache
- Close and reopen Slack app
"Missing Permissions" Error
Symptoms: Bot responds but says it can't perform action
Potential causes:
- Missing OAuth scopes
- Channel-specific permissions
- Private channel the bot can't access
Fix:
- Go to api.slack.com/apps ā Your App ā OAuth & Permissions
- Add the required scope (error message tells you which)
- Reinstall the app (required after scope changes)
- Verify bot is invited to the channel
Socket Mode Connection Fails
Symptoms: Logs show "Socket Mode connection failed" or "WebSocket error"
Potential causes:
- Firewall blocking WebSocket connections
- Proxy interfering
- Invalid app-level token
- Socket Mode not enabled
Fix:
# Test WebSocket connectivitycurl -i -N -H "Connection: Upgrade" \ -H "Upgrade: websocket" \ https://wss-primary.slack.com/# Verify app tokenecho $SLACK_APP_TOKEN # Should start with xapp-# Check Socket Mode is enabled in Slack app settings# Settings ā Socket Mode ā Should be ONMessages Sent But Not Received
Symptoms: Bot sends messages but they don't appear in Slack
Potential causes:
- Wrong channel ID
- Bot removed from channel
- Slack API rate limiting
- Invalid auth token
Fix:
# Test sending a messagecurl -X POST https://slack.com/api/chat.postMessage \ -H "Authorization: Bearer xoxb-your-token" \ -H "Content-Type: application/json" \ -d '{ "channel": "C01234567", "text": "Test message" }'# Check response for errors# If "channel_not_found" - verify channel ID# If "not_in_channel" - invite bot to channelHigh Latency (Slow Responses)
Symptoms: Bot takes 10+ seconds to respond
Potential causes:
- OpenClaw server overloaded
- Slow model inference
- Network latency
- Large context processing
Fix:
# Check OpenClaw resource usagetop -p $(pgrep openclaw)# Check model latencyopenclaw model benchmark# Enable response streaming for faster feel# In config.yaml:channels: slack: streaming: true # Send partial responses as they're generated"Rate Limited" Errors
Symptoms: Logs show "rate_limited" or "retry_after"
Potential causes:
- Too many messages sent too fast
- Slack's tier 2 rate limit (1 message per second per channel)
- Too many API calls
Fix:
- OpenClaw handles rate limiting automatically
- If seeing this often, you're sending too many messages
- Consider batching updates
- Use threads instead of new messages
# Configure stricter rate limitingchannels: slack: rateLimit: messagesPerSecond: 1 burstAllowance: 5Bot Goes Offline Randomly
Symptoms: Bot works, then stops, then works again
Potential causes:
- OpenClaw process crashing
- Socket Mode connection dropping
- Server restarting
- Memory issues
Fix:
# Check if OpenClaw is runningps aux | grep openclaw# Check for crashes in logsopenclaw gateway logs | grep -i error# Set up process monitoringsudo systemctl enable openclawsudo systemctl start openclaw# Or use PM2 for Node.js processespm2 start openclawpm2 savepm2 startupFile Uploads Not Working
Symptoms: Upload a file, bot doesn't acknowledge it
Potential causes:
- Missing
files:readscope - File too large
- File type not supported
- OpenClaw storage full
Fix:
- Add
files:readandfiles:writescopes - Reinstall app
- Check file size limits in config
- Ensure OpenClaw has disk space
Security Best Practices for Self-Hosted
1. Protect Your Tokens
# Never commit tokens to gitecho "config.yaml" >> .gitignore# Use environment variablesexport SLACK_BOT_TOKEN="xoxb-..."export SLACK_APP_TOKEN="xapp-..."# Reference in configchannels: slack: botToken: ${SLACK_BOT_TOKEN} appToken: ${SLACK_APP_TOKEN}# Or use secrets managementchannels: slack: botToken: "op://Work/Slack Bot/token" # 1Password reference2. Use Socket Mode
Socket Mode avoids exposing OpenClaw to the internet:
- No public URL needed
- No SSL certificate to manage
- No firewall rules to configure
- Reduced attack surface
3. Implement Access Controls
channels: slack: # Only allow specific users allowedUsers: - "U01234567" # You - "U09876543" # Your team lead # Only specific channels allowedChannels: - "C01234567" # #ai-help # Block sensitive channels ignoredChannels: - "C11111111" # #leadership - "C22222222" # #hr-confidential4. Enable Audit Logging
logging: audit: enabled: true logFile: "/var/log/openclaw/audit.log" includeMessages: false # Don't log message content (privacy) includeMetadata: true # Log who, when, where5. Rotate Tokens Regularly
- Every 90 days, regenerate Slack tokens
- Go to api.slack.com/apps ā Your App ā OAuth & Permissions
- Click "Rotate Token"
- Update OpenClaw config
- Restart OpenClaw
6. Monitor for Abuse
channels: slack: monitoring: alertOnHighUsage: true threshold: 100 # requests per hour notifyChannel: "C-SECURITY-ALERTS"7. Principle of Least Privilege
Only add OAuth scopes you actually need:
- Don't add
files:writeif you only read files - Don't add
groups:historyif you don't support private channels - Review scopes quarterly and remove unused ones
Comparison: HeraClaw Cloud vs Self-Hosted
| Feature | HeraClaw Cloud ⨠| Self-Hosted š ļø |
|---|---|---|
| Setup time | 60 seconds | 30-45 minutes (first time) |
| Technical skill | None | Intermediate to Advanced |
| Infrastructure | Managed for you | You maintain servers |
| Slack app config | Pre-configured | Manual OAuth setup |
| Updates | Automatic | Manual updates required |
| Slack API changes | We handle it | You must update |
| Support | Professional support team | Community forums |
| Uptime SLA | 99.9% guaranteed | DIY |
| Security certs | SOC 2, GDPR, HIPAA | Self-certification |
| Monitoring | 24/7 monitoring | You set up alerts |
| Cost | Monthly subscription | VPS + your time |
| Scaling | Auto-scales | Manual scaling |
| Multi-workspace | Unlimited | Limited by infra |
| Slash commands | Auto-registered | Manual registration |
| SSL/TLS | Included | You manage certificates |
| Data residency | US/EU regions | Your choice |
| Custom integrations | Request features | Full control |
| Compliance | We handle audits | You handle audits |
Why 95% of users choose HeraClaw Cloud:
ā Time is valuable - 30-45 min setup + ongoing maintenance adds up ā Slack API changes frequently - We handle breaking changes automatically ā Security is hard - SOC 2 compliance takes months and $$$ to achieve ā Reliability matters - 99.9% SLA vs DIY uptime monitoring ā Support is worth it - Email/chat/phone support when things break ā Hidden costs - SSL certs, monitoring tools, incident response time
When self-hosting makes sense:
ā You have DevOps engineers with spare capacity ā Data cannot leave your infrastructure (compliance) ā You're already running Kubernetes/Docker setups ā You want to heavily customize the codebase ā You have 24/7 on-call rotation to handle incidents
Frequently Asked Questions
Q: Do I need Slack Pro or Enterprise?
A: No. The integration works on Slack Free, Pro, Business+, and Enterprise Grid. Free tier has limited message history (90 days), but the bot works fine.
Q: Can I use this in multiple Slack workspaces?
A: Yes! HeraClaw Cloud supports unlimited workspaces. Self-hosted requires configuring each workspace separately.
Q: What permissions does the bot need?
A: Minimum: Send messages, read channel history, read DMs, respond to @mentions. Optional: Read files, add reactions. Full list in the setup guide above.
Q: Can I restrict who can use the bot?
A: Yes. Configure role-based access (only certain Slack roles), user-based access (specific users), or channel-based (only in certain channels).
Q: Does it work in private channels?
A: Yes, if you invite the bot to the private channel and grant the groups:history scope.
Q: Can it work in Slack Connect channels?
A: Yes, but the external organization must approve the bot. Both organizations must consent.
Q: What about Slack threads?
A: Full support. The bot understands thread context, responds in threads, and can summarize thread conversations.
Q: Is my data secure?
A: HeraClaw Cloud: SOC 2 Type II certified, encrypted in transit (TLS 1.3) and at rest (AES-256), GDPR compliant. Self-hosted: Your responsibility, but you control the infrastructure.
Q: Can the bot send proactive messages?
A: Yes! Schedule daily briefings, send alerts, or trigger messages based on events (GitHub PR created, customer signup, etc.).
Q: What if Slack changes their API?
A: HeraClaw Cloud: We handle all API changes automatically with zero downtime. Self-hosted: You must update OpenClaw manually when breaking changes occur.
Q: How fast are responses?
A: Typically 1-3 seconds from @mention to response, depending on:
- Model speed (GPT-4 vs Claude Sonnet vs local Llama)
- Context size (longer conversations = slower)
- Query complexity (simple Q&A vs multi-step reasoning)
Q: Can I use custom slash commands?
A: Yes. Create custom commands like /deploy, /incident, /search-docs that map to specific workflows.
Q: Does it support slash command auto-completion?
A: Yes, Slack's native auto-completion works. As you type /ask, Slack shows the command format.
Q: Can I migrate from self-hosted to Cloud?
A: Absolutely. Export your configuration, sign up for HeraClaw Cloud, import your settings. We'll help you migrate in under 15 minutes. Most users who migrate never go back.
Q: What about message rate limits?
A: Slack limits bots to ~1 message per second per channel. OpenClaw handles this automatically by queueing messages. You won't hit rate limits under normal usage.
Q: Can it mention users in responses?
A: Yes, the bot can @mention users if configured (requires users:read scope).
Q: Can I use emoji reactions?
A: Yes. The bot can react to messages with emoji (requires reactions:write scope). Useful for acknowledging requests.
Q: What about GIF and image support?
A: The bot can:
- Analyze uploaded images (describe, extract text, etc.)
- Send images (charts, diagrams generated by AI)
- Send GIFs (via URL)
Q: Can it integrate with other Slack apps?
A: Yes. If you use Jira, GitHub, Google Drive, or other Slack-integrated tools, OpenClaw can reference that data (e.g., "Show me my open Jira tickets").
Q: Does it work with Slackbot keywords?
A: OpenClaw is a separate bot, not Slackbot. But you can configure OpenClaw to respond to custom keywords.
Q: What about Enterprise Grid Shared Channels?
A: Supported. The bot can operate in shared channels across multiple Grid organizations (requires approval from all orgs).
Q: Can I customize the bot's name and avatar?
A: Yes! In Slack app settings, customize:
- Bot display name
- Bot username (@handle)
- Bot avatar/icon
- Bot description
Q: How do I remove the bot?
A: In Slack: Settings & Administration ā Manage Apps ā Find your bot ā Remove. Then disable in HeraClaw or OpenClaw config.
Get Started
For 95% of users (recommended):
- 60-second setup
- No technical skills needed
- Professional support
- 99.9% uptime SLA
- Try it risk-free
For advanced users only:
- 30-45 minute setup
- Requires DevOps skills
- You maintain infrastructure
- Full control and customization
Questions? Email support@cloud.getopenclaw.ai or join our community Slack for help.
Last updated: April 9, 2026
Built with OpenClaw ā The open-source AI assistant platform. Self-host or use HeraClaw Cloud.
Features
@mention support
Mention your assistant anywhere in Slack for instant help
Slash commands
Custom /commands for structured AI interactions across your workspace
Channel permissions
Configure exactly which channels your assistant operates in
Thread responses
Replies in threads to keep channels clean and organized
DM support
Private AI conversations via Slack direct messages
Enterprise OAuth
Secure workspace authentication with full Slack admin control
Use Cases
Engineering support
Code questions, PR reviews, and debugging help without leaving Slack
Sales enablement
CRM lookups, customer context, and response drafts on demand
Team knowledge base
Instant answers from wikis and docs embedded in workflow
Daily briefings
Automated morning updates posted to team channels
Setup Guide
Requirements
- āSlack workspace with app installation permissions
- āWorkspace Owner or Admin role (for custom app installation)
- āOpenClaw running on a server, VM, or Mac
- āFor Enterprise Grid: Org Owner or Admin role
Create a Slack App
Go to api.slack.com/apps and click 'Create New App'. Choose 'From scratch' and select your target workspace. Name your app (e.g., 'OpenClaw Assistant').
Enable Socket Mode
Under 'Socket Mode' in the left sidebar, enable it. This allows secure, real-time communication without exposing a public endpoint. Generate an App-Level Token with 'connections:write' scope.
Configure Bot Token Scopes
Under 'OAuth & Permissions', add Bot Token Scopes: app_mentions:read, channels:history, channels:read, chat:write, files:read, groups:history, groups:read, im:history, im:read, im:write, users:read. For slash commands, add commands scope.
Set up Event Subscriptions
Under 'Event Subscriptions', enable events and subscribe to: app_mention, message.channels, message.groups, message.im. This lets your bot receive messages.
Create Slash Commands (optional)
Under 'Slash Commands', create commands like /ask, /task, or /summarize. Set the Request URL to your Socket Mode handler (handled automatically by OpenClaw).
Install to Workspace
Under 'Install App', click 'Install to Workspace'. Authorize the requested permissions. Save the Bot User OAuth Token (xoxb-...) that's generated.
Configure OpenClaw
Add your Slack tokens to the OpenClaw config file. You'll need both the Bot Token (xoxb-...) and the App Token (xapp-...) from Socket Mode.
Test the Connection
Start OpenClaw with 'openclaw gateway start'. Invite your bot to a channel with /invite @YourBotName, then @mention it to verify responses.
Configuration Example
slack:
botToken: \"xoxb-your-bot-token\"
appToken: \"xapp-your-app-token\"
# Optional: Restrict to specific channels
allowedChannels:
- \"C0123456789\" # #ai-assistant
- \"C9876543210\" # #engineering
# Optional: Enable slash commands
slashCommands:
- command: \"/ask\"
description: \"Ask the AI assistant a question\"
- command: \"/summarize\"
description: \"Summarize a thread or document\"Limitations
- ā ļøWorkspace admin approval required for custom app installation
- ā ļøMessage history access requires paid Slack plans for full history
- ā ļøRate limits apply (varying by Slack plan tier)
- ā ļøSlack Connect channels require external org approval
Frequently Asked Questions
Do I need admin access to install?
Yes, installing custom Slack apps requires Workspace Owner or Admin permissions. In organizations with stricter policies, you may need to submit an app approval request through your IT team.
How does Enterprise Grid deployment work?
For Enterprise Grid, install the app at the organization level through your Org Admin Dashboard. You can then deploy to specific workspaces or enable org-wide. Each workspace inherits org-level permissions while allowing workspace-specific customization.
Does it work with SSO/SAML?
Yes. OpenClaw respects your Slack workspace's authentication settings. When users are authenticated via your SSO provider (Okta, Azure AD, OneLogin, etc.), the assistant recognizes their identity and can enforce role-based permissions.
What's the difference between channels and DMs?
In channels, the bot responds only when @mentioned or when slash commands are used. In DMs (direct messages), every message goes to the assistant. You can configure which mode works best for your team.
Can it work with Slack Connect channels?
Yes, but both organizations must approve. Your app needs to be installed in your workspace, and the external org must allow third-party apps in Connect channels. Check with both workspace admins.
Is it compliant with enterprise security requirements?
OpenClaw runs on infrastructure you control ā your servers, your cloud, your data policies. Messages are processed locally and never stored on third-party services. This architecture supports SOC 2, HIPAA, GDPR, and other compliance frameworks.
How do slash commands work?
Slash commands provide structured entry points like /ask [question] or /task [description]. They appear in Slack's autocomplete, making discovery easy. Configure them in your Slack app settings and OpenClaw handles the responses.
Can I restrict which channels the bot works in?
Yes. Set allowedChannels in your OpenClaw config to whitelist specific channel IDs. The bot will only respond in those channels, ignoring mentions elsewhere.
What about rate limits?
Slack enforces rate limits that vary by API method and workspace plan. OpenClaw handles rate limiting automatically with exponential backoff. For high-volume deployments, Enterprise Grid offers higher limits.
Can multiple team members configure the bot?
App installation and configuration requires admin access. However, you can build admin interfaces in App Home to let authorized users adjust settings without accessing the server config.
How does the App Home work?
App Home appears in users' Slack sidebar under 'Apps'. It provides a private space for each user to interact with the assistant, view their history, and access personalized settings or dashboards you configure.
Does it support message threads?
Yes, and it's the recommended interaction pattern. Responses are posted as thread replies to keep channels organized. The assistant maintains full context within a thread for multi-turn conversations.
š„ Your AI should run your business, not just answer questions.
We'll show you how.Free to join.
Related Integrations
š Learn More
Proxy, SSL & Network Connection Issues
Troubleshoot network connectivity, webhook integrations, Copilot Proxy setup, API credential issues, and connection timeouts.
OpenClaw Self-Audit: The Prompt That Checks Your Own Security Setup
Most people set up OpenClaw, lock things down as best they can, and never check again. Meanwhile, configs drift, updates change settings, and new integration...
How to Connect Gmail to Your AI Assistant (Complete Integration Guide)
Turn Gmail into an AI-powered inbox. Summarize threads, draft replies, auto-categorize messages, and search your email with natural language.
Tabnine vs GitHub Copilot
Privacy-first AI coding vs ecosystem integration
š Your AI should run your business.
Weekly live builds + template vault. We'll show you how to make AI actually work.Free to join.
Join Vibe Combinator ā