Cowork scouts & scaffolds. OpenClaw debugs & deploys. 97 AI use cases get built while you sleep β without burning $300/mo on OpenClaw doing everything alone.
AI USE CASES
GAME CHANGERS
STILL IDEAS
BUSINESSES TO TRANSFORM
OpenClaw is powerful but expensive. Using it for PRD writing, scaffolding, AND deployment means API costs spiral fast. We need a smarter split.
Included in Pro sub. Notion MCP, skills system, file creation, web research. $0 marginal cost.
Full repo access, debugging, deployment, IDE integration. Included with limits.
24/7 always-on, cron jobs, Telegram, messaging. But API costs = $30β300/mo.
KEY INSIGHT: Use each tool for what it's cheapest and best at. Cowork does the thinking & writing. OpenClaw does the 24/7 monitoring & coordination. Claude Code does the actual building.
| TASK | COWORK | CODE VS | OPENCLAW | BEST PICK |
|---|---|---|---|---|
| Write PRD / Spec | β Best | β οΈ Can | β οΈ Limited | COWORK |
| Notion DB Read/Write | β Native MCP | β οΈ Via MCP | β οΈ GOG skill | COWORK |
| File Creation (docx/pptx/pdf) | β Skills | β οΈ Raw code | β None | COWORK |
| Web Research | β Built-in | β οΈ Via MCP | β Brave API | COWORK |
| Scaffold React/Next.js | β No git | β Best | β οΈ Possible | CODE |
| Debug & Fix Code | β No repo | β Best | β οΈ Has terminal | CODE |
| Deploy (Vercel/Netlify) | β Cannot | β Best | β οΈ API calls | CODE |
| API Integration (Meta/TikTok) | β No env | β Best | β οΈ Possible | CODE |
| n8n Workflow Build | β Via MCP | β οΈ Via MCP | β οΈ API | COWORK |
| 24/7 Cron / Heartbeat | β Session dies | β No built-in | β Best | OPENCLAW |
| Telegram/Slack Alerts | β None | β None | β 15+ platforms | OPENCLAW |
| Proactive Monitoring | β Reactive only | β Reactive | β Reaches out first | OPENCLAW |
| Schedule/Coordinate | β οΈ Session-based | β None | β Best | OPENCLAW |
Each stage uses the cheapest, most capable tool for the job.
1. Read AI USE CASE DB (97 projects)
2. Filter: Game Changer + High impact, DVC/CC
3. Write full PRD inside each Notion page
4. Research competitors, APIs, tech stack
5. Generate scaffold spec (files, routes, schema)
6. Update status β "Planned" + tag "Ready for Code"
COST: $0 β included in Cowork Pro
1. Read PRD from Notion page
2. Scaffold project (Next.js/React/Python)
3. Implement core features
4. Debug autonomously (test β fix loop)
5. Deploy to Vercel/Netlify
6. Update Notion status β "Done" + deploy URL
COST: Included β Claude Code limits
1. Monitor deployed apps 24/7
2. Run scheduled cron jobs
3. Send Telegram alerts on errors
4. Coordinate handoffs between stages
5. Overnight research for next projects
6. Morning briefing via Telegram
COST: ~$15-30/mo β lightweight tasks only
Sorted by: feasibility (easy first) Γ business impact. DVC and Content Clinic projects only.
Three scheduled tasks that keep the factory humming, each on the right platform.
OpenClaw checks: AI USE CASE DB for "Ready for Code" projects β queues them in a build list β sends Telegram morning briefing to Fasai with today's build targets. Also checks deployed apps health + overnight errors.
Runs on schedule or on-demand: queries AI USE CASE DB β picks next "Idea" project (Game Changer/High, DVC/CC) β writes full PRD with tech stack, API research, component specs, database schema β updates status to "Planned" β adds "Ready for Code" tag. One project per run. Batch of 3-5 per day.
Fasai opens Code VS β picks a "Planned" project with full PRD β Claude Code scaffolds, builds, tests, deploys β updates Notion status to "Done" or "Showcase" with deploy URL. OpenClaw notified of completion.
OpenClaw runs health checks on all deployed apps β researches APIs for tomorrow's builds (Meta, Pancake, TikTok docs) β pre-writes research notes into Notion β queues next day's spec targets. Sends summary if anything breaks.
A scheduled Cowork task that turns Ideas β fully-specced Planned projects, ready for Code VS.
Query AI USE CASE DB β filter Impact = Game Changer or High, Business = DVC or CC, Status = Idea
Pick the highest-impact unspecced project β prioritize Easy feasibility first, then Medium
Web research β search for relevant APIs, competitor tools, open-source options
Write full PRD inside the Notion page β problem, solution, tech stack, API endpoints, DB schema, component tree, deployment plan
Update properties β Status β Planned, Recommended Platform, Tools Needed, Skills Needed, Auth Required
Log to Session Log β what was specced, any blockers found, next steps
β 1 project fully specced with PRD
β Status updated to "Planned"
β All DB properties filled
β Session log entry created
β Ready for Claude Code to build
3Γ per day β morning, afternoon, evening runs
Or on-demand via Fasai trigger
~48 ideas β Planned in 2 weeks
VPS hosting: $5β12/mo
API for PRD writing: ~$50-80/mo
API for coding/debug: ~$100-150/mo
API for monitoring: ~$30-50/mo
Cowork PRDs + specs: $0 (Pro sub)
Code VS build/debug: $0 (included)
OpenClaw VPS: $5-12/mo
OpenClaw API (monitoring only): ~$15-30/mo
Easy + Medium feasibility Game Changers and High-impact projects. These ship first.
| # | PROJECT | BUSINESS | IMPACT | STAGE 1 (COWORK) | STAGE 2 (CODE) | STAGE 3 (CLAW) |
|---|---|---|---|---|---|---|
| 1 | Brand & Product Marketing Planner | DVC | High | PRD + strategy framework | React dashboard | β |
| 2 | Cost Breakdown & Feasibility Study | DVC | High | PRD + cost model | Calculator app | β |
| 3 | Product Info Writer (from Images) | DVC | Medium | Already has skill β | Web wrapper | β |
| 4 | Video Auto Transcriber | CC | High | PRD + Whisper research | Python pipeline | Cron trigger |
| 5 | Customer Audit & Research Agent | CC | High | Already has skill β | API wrapper | β |
| 6 | Sales Promotion Planner | DVC | High | PRD + promo logic | Web app | Schedule alerts |
| 7 | Best Post Ad Runner | DVC | Game Changer | PRD + Meta API spec | Meta API integration | 24/7 monitor |
| 8 | Personal Brand Proposal Generator | CC | Game Changer | Already has skill β | Notion auto-trigger | β |
| 9 | Monthly Report Generator | CC | Game Changer | Already has skill β | Auto-pull data | Scheduled run |
| 10 | SEO Website Generator | CC | Game Changer | Already has skill β | Deploy pipeline | Health monitor |
Only the things no other tool can do. Maximum value, minimum API tokens.
Check AI USE CASE DB every 6 hours for "Ready for Code" projects. Queue them. Track progress. Ping Fasai on Telegram when a project is ready for Code VS session.
Morning: today's build queue + overnight status. Evening: what shipped, what's blocked. Alert: if any deployed app errors. This is OpenClaw's #1 unique value.
Research APIs for tomorrow's builds: Meta Graph API, Pancake docs, TikTok Business API. Pre-write findings into Notion so Cowork/Code have context ready.
Ping deployed apps (GeoScout, CEO websites, dashboards). Check if they're up. Alert via Telegram if anything is down. Log incidents.
Check FJ Project Tasks DB. Notify Fasai when Mix updates tasks. Coordinate handoffs between Fasai's builds and Mix's Bubble.io work.
No PRD writing. No file creation. No coding. No debugging. No deployment. Those are Cowork and Code VS jobs now. OpenClaw stays lean.
β’ Add "Ready for Code" to Workflow property in AI USE CASE DB
β’ Create Cowork scheduled task: "Spec Factory" β queries Idea projects, writes PRDs
β’ Test with 2-3 Easy/DVC projects first
β’ Verify PRD quality before scaling
β’ Create "OpenClaw Bot" Notion integration
β’ Share AI USE CASE + Session Log DBs
β’ Build OpenClaw skill: "build-queue-scanner"
β’ Set up Telegram morning briefing cron
β’ Run Spec Factory 3Γ daily on schedule
β’ All 48 "Idea" projects get PRDs in ~2 weeks
β’ Start daily Code VS build sessions (1-2 projects/day)
β’ OpenClaw tracks completion + sends updates
β’ Pipeline runs end-to-end: Idea β Planned β Built β Deployed
β’ OpenClaw overnight research feeds next day's specs
β’ Fasai's only job: approve and test during Code VS sessions
β’ Target: 5-8 projects shipped per week
PER MONTH
(down from $185-292)
IDEA BACKLOG CLEARED
in 2-3 weeks
PROJECTS SHIPPED
per week at full speed
NEXT STEP: Create the Cowork "Spec Factory" scheduled task today. Start with 3 Easy DVC projects. Prove the pipeline works before scaling.