Recipes
Recipes
Practical setups that extend your OpenClaw deployment beyond the basics. Each recipe assumes you’ve completed the relevant phase guides and layers on additional configuration, AGENTS.md instructions, and cron jobs.
Recipes
| Recipe | What it does | Prereqs |
|---|---|---|
| Knowledge Vault | Persistent knowledge base with semantic search and cron-driven autonomous research | Phases 1–5, search agent |
| Morning Briefing | Scheduled daily briefing — weather, calendar, inbox, tasks, and news — delivered to your channel | Phases 1–5, search agent |
Use Case Ideas
Not every use case warrants a full recipe page. The following are patterns with strong community adoption — enough to get started with the right AGENTS.md instructions and cron jobs on top of the phase guides.
Personal Productivity
| Use Case | What it does | Key tools |
|---|---|---|
| Personal CRM | Auto-track contacts from Gmail + Calendar, filter newsletters/bots, score by interaction strength, queryable via chat | gog skill, memory_search, SQLite via exec, cron |
| Meeting → task extraction | Parse transcripts (Fathom, Otter.ai) → extract action items → approval gate → create in Todoist/Linear | web_fetch, memory, approval in AGENTS.md |
| Smart inbox triage | Flag urgent email, summarize newsletters into a digest, auto-reply to known senders | gog skill, cron, AGENTS.md triage rules |
| Second brain | Text anything via chat → stored in memory; agent builds a searchable Next.js dashboard on request | Memory tools, exec, Telegram/iMessage |
| Family calendar assistant | Aggregate family calendars → morning briefing; monitor messages for appointment confirmations; pantry inventory from photos | gog skill, image tool, cron, iMessage |
Development & DevOps
| Use Case | What it does | Key tools |
|---|---|---|
| GitHub PR review | New PR → fetch diff → analyze → notify developer via channel (read-only: gh pr view, gh run view) | exec (gh CLI), message, cron or webhook |
| Self-healing home server | Hourly health checks (Gatus, ArgoCD); auto-restart failing services; page only when auto-remediation fails | exec (kubectl/ssh), cron, message |
| CI/CD monitor | Watch build status, detect regressions, post summaries to Slack/Discord | exec, web_fetch, cron, message |
| Overnight mini-app builder | Brain-dump goals → agent generates + executes daily tasks, including building a small app overnight | exec, browser, sessions_spawn, cron |
Multi-Agent Orchestration
| Use Case | What it does | Key config |
|---|---|---|
| Hub-and-spoke team | Orchestrator agent delegates to specialized workers (Coder, Researcher, Automation). Workers run in Docker with minimal tools | Multiple agents in agents.list, subagents.allowAgents, Docker sandbox |
| Autonomous project management | Subagents coordinate via a shared STATE.yaml file. Main agent stays thin (strategy only). Workers self-assign, update state, and report back | sessions_spawn, group:fs, AGENTS.md delegation rules |
| Multi-agent content factory | Research, writing, and thumbnail agents in separate Discord/Slack channels, each specialized and channel-bound | Bindings, multiple agents, subagents.allowAgents |
Business & Analytics
| Use Case | What it does | Key tools |
|---|---|---|
| Nightly business briefing | Multi-persona AI council (GrowthStrategist, RevenueGuardian, SkepticalOperator) reviews business signals and produces ranked recommendations | sessions_spawn, cron, AGENTS.md persona definitions |
| Competitor monitoring | Daily web search on competitor names + products; append findings to a knowledge file; alert on significant changes | web_search via search agent, cron, memory |
| Multi-channel customer service | 24/7 auto-responses across WhatsApp, Instagram, Email, Google Reviews with escalation rules | Channel bindings, AGENTS.md response rules, message |
| YouTube analytics | Track channel metrics, monitor competitor videos, generate weekly performance charts | web_fetch, exec, cron, memory |
| Dynamic dashboard | Cron every 15 min spawns parallel sub-agents for GitHub/social/server metrics; aggregates to a database | sessions_spawn, exec, PostgreSQL/SQLite |
Research & Intelligence
| Use Case | What it does | Key tools |
|---|---|---|
| Multi-source tech news digest | 109+ sources (46 RSS + 44 Twitter/X KOLs + 19 GitHub repos + web search), quality-scored and deduplicated | tech-news-digest
ClawHub skill |
| Market research → MVP factory | Mine Reddit + X for pain points in any niche (Last 30 Days skill) → rank opportunities → build MVP | last-30-days
skill, exec, browser |
| AI earnings tracker | Track tech/AI earnings reports with automated previews, alerts, and summaries | web_search via search agent, cron, memory |
Infrastructure & Integrations
| Use Case | What it does | Key config |
|---|---|---|
| n8n credential isolation | Agent calls n8n webhooks for all external API interactions — API keys live in n8n only, never in agent environment. Visual, lockable pipelines | web_fetch (webhook calls), Docker Compose (openclaw-n8n-stack
), AGENTS.md workflow conventions |
| Phone-based assistant | Access your agent via phone call or SMS — calendar, Jira, web search hands-free | ClawdTalk plugin (Telnyx) |
| AI voice calls (batch) | Call a list of people using an AI persona — guest confirmations, appointment reminders | SuperCall plugin (Twilio + GPT-4o Realtime) |
Home Automation
| Use Case | What it does | Key tools |
|---|---|---|
| Home Assistant control | Natural language device control, automations, and status queries via any channel | Home Assistant add-on, web_fetch (HA REST API), message |
| Smart energy monitoring | IoTawatt calibration alerts, anomaly detection, usage reports | web_fetch, cron, memory |
| EV charge scheduling | Tesla/EV charge scheduling and status queries via chat | web_fetch (Tesla API or TeslaMate), message |
Contribute a recipe. If you’ve built something that works reliably, the community collection at awesome-openclaw-usecases
accepts pull requests. Each entry needs a working setup you’ve tested for at least a day.
Last updated on