Daisy turns raw customer feedback into prioritized roadmaps and sprint-ready development tasks. It automates the entire product management workflow — from ingesting data sources (CRM, support tickets, surveys) to extracting insights, recommending features, writing specs, and breaking work into developer tasks.
Built at TreeHacks 2025.
- Data Source Integration — Connect tools like Salesforce, Zendesk, Intercom, Slack, Gong, HubSpot, Jira, Linear, and Amplitude, or upload custom documents (interviews, tickets, surveys).
- AI-Powered Insight Extraction — Automatically categorize customer feedback into pain points, feature requests, praise, and confusion with severity scoring, frequency tracking, and direct customer quotes.
- Feature Recommendations — Generate prioritized feature suggestions scored by impact, effort, confidence, and overall priority.
- Spec Generation — One-click PRD and technical spec creation including user stories, UI changes, data model changes, and success metrics.
- Task Breakdown — Convert features into development tasks with effort estimates, acceptance criteria, and task types.
- Roadmap Management — Sprint-based roadmap with drag-and-drop assignment, story point tracking, team member assignment, and status tracking.
- RAG Chatbot (Daisy) — Context-aware AI assistant with full board knowledge that can answer questions about sources, insights, features, specs, tasks, and the roadmap.
| Layer | Technology |
|---|---|
| Framework | Next.js 14 (App Router) |
| UI | React 18, Tailwind CSS, Framer Motion |
| State | Zustand (with localStorage persistence) |
| Visualization | React Flow |
| LLM | Anthropic Claude / OpenAI / xAI Grok (configurable) |
| Streaming | Server-Sent Events (SSE) |
treehacks/
├── app/
│ ├── api/
│ │ ├── analyze/route.js # Sources → Insights
│ │ ├── recommend/route.js # Insights → Feature recommendations
│ │ ├── spec/route.js # Features → Technical specs
│ │ ├── tasks/route.js # Features → Development tasks
│ │ ├── chat/route.js # RAG chatbot endpoint
│ │ └── research/route.js # Research functionality
│ ├── layout.js
│ └── page.js
├── components/
│ ├── ChatPanel.jsx # Floating chat interface
│ ├── Dashboard.jsx # Main dashboard
│ ├── FeatureDetail.jsx # Feature detail view
│ ├── FeaturesView.jsx # Feature list
│ ├── InsightsView.jsx # Insights list
│ ├── IntegrationLogos.jsx # Integration icons
│ ├── LandingPage.jsx # Marketing landing page
│ ├── RoadmapView.jsx # Sprint roadmap
│ ├── Sidebar.jsx # Navigation sidebar
│ └── SourcesView.jsx # Integrations & sources
├── lib/
│ ├── agents/
│ │ ├── analyzer.js # Insight extraction agent
│ │ ├── recommender.js # Feature recommendation agent
│ │ ├── specwriter.js # Spec generation agent
│ │ └── taskbreaker.js # Task breakdown agent
│ ├── research/
│ │ ├── stagehand.js # Browser automation research
│ │ ├── perplexity.js # Perplexity API integration
│ │ └── mock.js # Mock research data
│ └── llm.js # LLM adapter (multi-provider)
├── store/
│ └── useProductStore.js # Zustand store
├── data/ # Sample data & integration mocks
├── package.json
├── tailwind.config.js
└── next.config.js
- Node.js 18+
- An API key for one of the supported LLM providers (Anthropic, OpenAI, or xAI)
git clone https://github.com/your-username/treehacks.git
cd treehacks
npm installCreate a .env.local file in the project root:
# LLM provider: "anthropic", "openai", or "openai-compatible"
LLM_PROVIDER=openai-compatible
# Model name
LLM_MODEL=grok-4-1-fast-reasoning
# Base URL (required for openai-compatible providers)
LLM_BASE_URL=https://api.x.ai/v1
# API key for your chosen provider
LLM_API_KEY=your-api-key-herenpm run devnpm run build
npm startConnect Sources → Analyze → Recommend → Spec → Tasks → Roadmap
(CRM, docs) (insights) (features) (PRDs) (dev work) (sprints)
- Connect integrations or upload documents as sources.
- Analyze sources to extract categorized insights.
- Recommend features based on the insights, scored by impact and effort.
- Generate specs with user stories, technical details, and success metrics.
- Break down tasks with estimates and acceptance criteria.
- Plan sprints on the roadmap with drag-and-drop, assignments, and status tracking.
Use the Daisy chatbot at any step to ask questions about your board.
MIT