UNPKG

oneie

Version:

Build apps, websites, and AI agents in English. Zero-interaction setup for AI agents (Claude Code, Cursor, Windsurf). Download to your computer, run in the cloud, deploy to the edge. Open source and free forever.

702 lines (535 loc) 19.2 kB
--- title: Workflow Onboarding dimension: connections category: workflow-onboarding.md tags: ai, cycle, ontology related_dimensions: events, groups, people, things scope: global created: 2025-11-03 updated: 2025-11-03 version: 1.0.0 ai_context: | This document is part of the connections dimension in the workflow-onboarding.md category. Location: one/connections/workflow-onboarding.md Purpose: Documents one platform onboarding workflow Related dimensions: events, groups, people, things For AI agents: Read this to understand workflow onboarding. --- # ONE Platform Onboarding Workflow **Version:** 1.0.0 **Purpose:** Complete user journey from `npx oneie init` to deployed website **Paradigm:** Cycle-based execution (not time-based) --- ## The Complete Journey ``` 1. Init (1 min) 2. Profile Setup (2 min) 3. AI Website Analysis (30 sec) 4. Custom Ontology Generation (30 sec) 5. Feature Selection (2 min) 6. Cycle Plan Generation (10 sec) 7. Landing Page Deploy (Cycle 1-10, ~5 min) 8. Full Site Build (Cycle 11-100, ~30-60 min) 9. Live Production Site ``` **Total Time:** ~45-75 minutes from zero to production --- ## Phase 1: Init & Profile Setup ### Step 1: `npx oneie init` ```bash $ npx oneie init ██████╗ ███╗ ██╗███████╗ ██╔═══██╗████╗ ██║██╔════╝ ██║ ██║██╔██╗ ██║█████╗ ██║ ██║██║╚██╗██║██╔══╝ ╚██████╔╝██║ ╚████║███████╗ ╚═════╝ ╚═╝ ╚═══╝╚══════╝ Make Your Ideas Real ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Welcome! Let's build your platform. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ? What's your name? Tom O'Connor ? Organization name? ONE Platform ? What's your current website? https://one.ie ? What email should we use? tom@one.ie Perfect! Starting Claude Code to analyze your website... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🤖 Launching Claude Code ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ $ claude Claude Code initialized. Running command: $ /one analyze https://one.ie --org "ONE Platform" --email tom@one.ie ``` **Creates:** - `.env.local` with user details - Installation folder: `/one-platform/` - Initial group structure ### Step 2: Website Analysis (Claude Code + agent-onboard) ```bash ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🤖 CLAUDE CODE ACTIVE ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Running: /one analyze https://one.ie agent-onboard is analyzing your website... [WebFetch] Fetching https://one.ie [WebFetch] Analyzing HTML, CSS, and content [WebFetch] Extracting brand identity Detected: Primary color: #FF6B6B (Coral) Secondary color: #4ECDC4 (Teal) Logo: /logo.svg Brand voice: Technical, friendly, empowering Target audience: Developers, creators, entrepreneurs Existing platform: Yes (Astro + Convex) Extracted features: Multi-tenant architecture AI agent integration Creator economy tools Real-time collaboration 6-dimension ontology Generated custom ontology: Groups: Organizations, teams People: Developers, creators Things: Projects, content, code Connections: Collaborations, integrations Events: Deployments, commits, releases Knowledge: Documentation, patterns [Write] Creating /one-platform/knowledge/ontology.md [Write] Creating /one-platform/knowledge/brand-guide.md [Write] Creating .env.local 📁 Installation folder created: /one-platform/ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ANALYSIS COMPLETE ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ``` **How it works:** 1. CLI runs `claude` to start Claude Code 2. CLI sends `/one analyze <url>` slash command 3. Claude Code invokes **agent-onboard** 4. agent-onboard uses WebFetch, Write, Edit tools 5. Results saved to installation folder 6. Control returns to CLI for feature selection ### Step 3: Custom Ontology Confirmation ```bash ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 📊 YOUR CUSTOM ONTOLOGY ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Based on https://one.ie, I've created a custom ontology: GROUPS (Who organizes): organizations, teams, projects PEOPLE (Who acts): platform_owner, org_owner, developer, creator THINGS (What exists): projects, websites, agents, content, features deployments, repositories, integrations CONNECTIONS (How they relate): owns, collaborates_on, deployed_to, integrated_with EVENTS (What happens): project_created, feature_deployed, agent_executed content_published, integration_connected KNOWLEDGE (What's learned): documentation, patterns, best_practices, lessons ? Does this match your vision? (y/n) y ``` --- ## Phase 2: Feature Selection ```bash ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ AVAILABLE FEATURES ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Based on your ontology, here are recommended features: FOUNDATION (Recommended) [x] Landing page (Cycle 1-10, ~5 min) [x] Authentication (Cycle 11-20, ~10 min) [x] Multi-tenant groups (Cycle 21-30, ~10 min) CREATOR FEATURES [ ] Content publishing (Cycle 31-40, ~15 min) [ ] Membership tiers (Cycle 41-50, ~15 min) [ ] Revenue tracking (Cycle 51-60, ~10 min) DEVELOPER FEATURES [ ] Project management (Cycle 31-40, ~15 min) [ ] Code repositories (Cycle 41-50, ~15 min) [ ] Deployment pipeline (Cycle 51-60, ~20 min) AI FEATURES [ ] AI agents (Cycle 61-70, ~20 min) [ ] RAG knowledge base (Cycle 71-80, ~15 min) [ ] Semantic search (Cycle 81-90, ~10 min) ? Select features to build: (Space to select) ``` ### Feature Selection Output ```bash Selected Features: 1. Landing page (Cycle 1-10) 2. Authentication (Cycle 11-20) 3. Multi-tenant groups (Cycle 21-30) 4. Project management (Cycle 31-40) 5. AI agents (Cycle 61-70) 📊 Total Cycles: 50/100 ⏱️ Estimated Time: ~60 minutes 💰 Estimated Cost: $0 (free tier) ? Ready to start? (y/n) y ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🚀 STARTING BUILD ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Strategy: Landing page first, then iterate Step 1: Deploy landing page (Cycle 1-10) Step 2: Add authentication (Cycle 11-20) Step 3: Add multi-tenancy (Cycle 21-30) Step 4: Add project management (Cycle 31-40) Step 5: Add AI agents (Cycle 61-70) Step 6: Final deploy & docs (Cycle 91-100) ``` --- ## Phase 3: Landing Page Deploy (Cycle 1-10) ### Cycle Sequence ```bash 🤖 Starting Cycle 1-10: Landing Page MVP ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Cycle 1/10: Validate ontology and create installation folder Created /one-platform/ Created /one-platform/knowledge/ontology.md Validated 6-dimension structure Cycle 2/10: Map landing page to ontology Thing: "landing_page" entity Properties: { hero, features, cta, footer } Scoped to: groupId (ONE Platform) Cycle 3/10: Design landing page wireframe Hero: Logo, tagline, CTA Features: 3-column grid Footer: Links, social Colors: #FF6B6B, #4ECDC4 Cycle 4/10: Generate Astro page Created: /web/src/pages/index.astro Using: Brand colors, logo Components: Hero, Features, CTA Cycle 5/10: Add Tailwind styling Theme: Custom colors from website Responsive: Mobile-first Typography: From brand guide Cycle 6/10: Build and test locally Build: Success (0 errors) Lighthouse: 100/100/100/100 Test: Responsive design works Cycle 7/10: Deploy to Cloudflare Pages Build: Production ready Deploy: https://one-platform.pages.dev SSL: Enabled Edge: 330+ locations Cycle 8/10: Add custom domain DNS: CNAME configured Domain: https://one-platform.com SSL: Auto-provisioned Cycle 9/10: Test production deployment URL: Live and working Speed: <330ms globally SEO: Meta tags added Cycle 10/10: Document and notify Created: /one-platform/events/deploy-landing.md Screenshot: Saved to knowledge Email: Sent to tom@one.ie ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ LANDING PAGE LIVE! ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🌐 https://one-platform.com <330ms average response time 📱 100/100/100/100 Lighthouse scores 🌍 Live on Cloudflare Edge (330+ locations) Next: Add authentication (Cycle 11-20) Time: ~10 minutes ? Continue to next phase? (y/n) y ``` --- ## Phase 4: Iterative Feature Building ### Cycle 11-20: Authentication ```bash 🤖 Starting Cycle 11-20: Authentication ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Cycle 11: Map auth to ontology People: user accounts Events: signin, signup, signout Knowledge: Sessions, tokens Cycle 12-14: Backend schema & mutations Better Auth + Convex adapter Mutations: signup, signin, signout Queries: getSession, getUser Cycle 15-17: Frontend pages /account/signin /account/signup /account/settings Cycle 18-19: Test & deploy Tests: 15/15 passing Deploy: Live at https://one-platform.com/account Cycle 20: Document Updated: /one-platform/events/auth-deployed.md Authentication LIVE! ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Next: Multi-tenant groups (Cycle 21-30) ``` ### Cycle 21-30: Multi-Tenant Groups ```bash 🤖 Starting Cycle 21-30: Multi-Tenant Groups ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Cycle 21: Map groups to ontology Groups: Hierarchical (parent/child) Types: organization, team, project Data scoping: All entities groupId Cycle 22-25: Backend implementation Schema: groups table with parentGroupId Mutations: createGroup, updateGroup Queries: listGroups, getGroupHierarchy Cycle 26-28: Frontend dashboard /dashboard/groups /dashboard/groups/[id] Group switcher in nav Cycle 29: Test & deploy Tests: 25/25 passing Deploy: Live Cycle 30: Document Updated: /one-platform/events/groups-deployed.md Multi-Tenancy LIVE! ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Next: Project management (Cycle 31-40) ``` --- ## Phase 5: Specialized Agents ### Agent Coordination Each feature type gets handled by specialized agents: ```typescript const AGENT_MAPPING = { // Backend features "authentication": "agent-backend", "multi-tenant-groups": "agent-backend", "data-schema": "agent-backend", // Frontend features "landing-page": "agent-frontend", "dashboard": "agent-frontend", "user-profile": "agent-frontend", // Design features "wireframes": "agent-designer", "component-library": "agent-designer", "brand-guide": "agent-designer", // Integration features "stripe-payments": "agent-integrator", "discord-community": "agent-integrator", "email-marketing": "agent-integrator", // AI features "rag-knowledge": "agent-backend", // Convex vector search "semantic-search": "agent-backend", "ai-agents": "agent-builder", }; ``` --- ## Phase 6: Complete Build (Cycle 31-100) ### Example: Project Management Feature ```bash 🤖 Starting Cycle 31-40: Project Management ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Assigned to: agent-backend, agent-frontend Cycle 31: Map to ontology Things: projects, tasks, milestones Connections: assigned_to, depends_on Events: project_created, task_completed Cycle 32-34: Backend (agent-backend) Schema: things table extensions Mutations: createProject, updateTask Queries: listProjects, getProjectTasks Cycle 35-37: Frontend (agent-frontend) Pages: /projects, /projects/[id] Components: ProjectCard, TaskList Islands: Interactive task board Cycle 38: Tests (agent-quality) Unit: 20/20 passing Integration: 10/10 passing Cycle 39: Deploy (agent-ops) Backend: Convex deployed Frontend: Cloudflare deployed Cycle 40: Document (agent-documenter) Updated: /one-platform/events/projects-deployed.md Knowledge: Usage patterns captured Project Management LIVE! ``` --- ## Key Features of This Workflow ### 1. Immediate Value - **Landing page in 5 minutes** - User sees progress immediately - **Live URL from Cycle 10** - Something to share right away - **Iterative deployment** - Each feature goes live as it's built ### 2. AI-Driven Personalization - **Website analysis** - Extracts colors, brand, features automatically - **Custom ontology** - Generated from existing website/business - **Smart feature recommendations** - Based on detected patterns ### 3. Transparent Progress - **Cycle-based** - Clear progress (Cycle 23/100) - **Time estimates** - Per-feature and total - **Real-time updates** - See what's happening ### 4. Modular & Extensible - **Pick features** - Only build what you need - **Add later** - Resume from any cycle - **Custom features** - Extend with your own ontology ### 5. Multi-Agent Coordination - **agent-onboard** - Website analysis & ontology generation - **agent-director** - Feature planning & agent coordination - **agent-backend** - Convex schema, mutations, queries - **agent-frontend** - Astro pages, React components - **agent-designer** - Wireframes, brand system - **agent-quality** - Testing & validation - **agent-ops** - Deployment & monitoring - **agent-documenter** - Knowledge capture --- ## Implementation Steps ### 1. Create agent-onboard ```typescript // .claude/agents/agent-onboard.md # Agent: Onboard **Purpose:** Analyze existing website and generate custom ontology **Responsibilities:** 1. Fetch and analyze website URL 2. Extract brand colors, logo, fonts 3. Detect existing features and tools 4. Identify target audience and tone 5. Generate custom 6-dimension ontology 6. Create installation folder structure 7. Populate initial documentation **Tools:** WebFetch, Read, Write, Edit **Output:** - `/[installation-name]/knowledge/ontology.md` - `/[installation-name]/groups/[group-name]/` - `.env.local` with configuration ``` ### 2. Update agent-director ```typescript // Add to agent-director responsibilities: - Coordinate with agent-onboard for initial setup - Map user feature selections to cycle ranges - Generate 100-cycle plan based on selections - Assign features to appropriate specialists - Track progress and notify user ``` ### 3. Create CLI commands ```bash # cli/bin/oneie.js #!/usr/bin/env node import { Command } from 'commander'; import { init, build, deploy } from '../src/commands/index.js'; const program = new Command(); program .name('oneie') .description('ONE Platform CLI - Make Your Ideas Real') .version('1.0.0'); program .command('init') .description('Initialize new ONE platform') .action(init); program .command('build') .description('Build features using 100-cycle plan') .option('-f, --features <features...>', 'Features to build') .action(build); program .command('deploy') .description('Deploy to production') .action(deploy); program.parse(); ``` ### 4. Feature Library ```typescript // cli/src/features/library.ts export const FEATURES = { "landing-page": { name: "Landing Page", description: "Beautiful landing page with hero, features, CTA", cycles: [1, 10], duration: "~5 min", cost: "$0", requires: [], specialist: "agent-frontend", }, "authentication": { name: "Authentication", description: "Email/password + OAuth (6 methods)", cycles: [11, 20], duration: "~10 min", cost: "$0", requires: ["landing-page"], specialist: "agent-backend", }, "multi-tenant": { name: "Multi-Tenant Groups", description: "Hierarchical groups with data scoping", cycles: [21, 30], duration: "~10 min", cost: "$0", requires: ["authentication"], specialist: "agent-backend", }, // ... more features }; ``` --- ## Success Metrics - [ ] `npx oneie init` creates installation folder - [ ] agent-onboard extracts brand from website - [ ] Custom ontology generated automatically - [ ] Feature selection UI shows cycles + time - [ ] Landing page deploys in <10 minutes - [ ] Each feature goes live incrementally - [ ] User receives email after each phase - [ ] Full site built via 100-cycle plan --- ## Example End-to-End Flow ```bash # User runs init $ npx oneie init # Answers prompts ? Name: Tom O'Connor ? Org: ONE Platform ? Website: https://one.ie ? Email: tom@one.ie # AI analyzes website (30 seconds) Brand extracted Custom ontology created Features recommended # User selects features [x] Landing page (Cycle 1-10) [x] Authentication (Cycle 11-20) [x] Multi-tenant (Cycle 21-30) [x] Project management (Cycle 31-40) # Build starts automatically 🚀 Starting 40-cycle build... # 5 minutes later Landing page live: https://one-platform.com # 10 minutes later Authentication added: /account/signin # 20 minutes later Multi-tenancy working: /dashboard/groups # 35 minutes later Projects live: /projects # Final notification ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🎉 YOUR PLATFORM IS LIVE! ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🌐 https://one-platform.com 📱 /account - Authentication 🏢 /dashboard/groups - Organizations 📊 /projects - Project management 📊 Stats: Cycles: 40/100 complete Time: 35 minutes Cost: $0 Speed: <330ms globally Want to add more features? Run: $ npx oneie build --features ai-agents rag-knowledge 📚 Docs: https://docs.one.ie 💬 Support: https://discord.gg/one ``` --- **This is cycle-based onboarding done right!** 🚀 Fast. Transparent. Incremental. Personal. Perfect.