# Prisma

> The "what to build" IDE for product teams

Prisma is a desktop application for macOS that helps product managers discover what to build next. It connects to 6 tools (Slack, Linear, GitHub, PostHog, Notion, Jira), conducts AI-assisted customer interviews, synthesizes research from uploaded documents and recordings, and generates 9-section feature dossiers — evidence-backed specs ready for coding agents like Cursor, Claude Code, and GitHub Copilot to implement.

Your team ships code faster than ever. Prisma makes sure they ship the right thing.

## Core Workflow
1. Connect: Slack, Linear, GitHub, PostHog, Notion, Jira — 6 integrations via OAuth/API key
2. Collect: Auto-classify signals into 5 types — pain points, feature requests, bugs, praise, analytics insights
3. Research: Upload audio, video, PDFs, docs. Auto-transcribe with Whisper. AI synthesizes themes, pain points, feature requests, and data gaps
4. Interview: AI Copilot sits in customer calls — dual audio capture, real-time transcription, contextual follow-up questions, structured debriefs
5. Analyze: Canvas AI cross-references signals from all sources to find patterns humans miss
6. Decide: Generate a 9-section Feature Dossier — Problem, Research Summary, Solution, Technical Spec, UI/UX Spec, Impact Analysis, Risks, Alternatives, Acceptance Criteria

## Key Features
- Feature Dossier System: 9-section spec builder. AI generates each section from real signal evidence. Every claim links to specific customer quotes, analytics data, or ticket references. Output is structured for handoff to coding agents.
- Research Tab: Upload customer data (audio, video, PDF, docs). Auto-transcribe with Whisper. AI synthesis produces key themes, pain points, feature requests, positive feedback, and data gaps. "Generate Features" creates AI-proposed feature candidates from research.
- Interview Copilot: Real-time AI assistant for customer calls. Dual audio capture (interviewer + customer), real-time transcription, contextual follow-up suggestions. Writes structured debriefs. Turns insights into signals automatically.
- AI Canvas: Visual workspace to organize signals, run Analyze mode (cross-source pattern detection), and generate feature dossiers from signal clusters. Supports 11 node types including features, insights, research, and integration nodes.
- Signal Aggregation: 5 signal types (pain_point, feature_request, bug_report, praise, analytics_insight) from 6 sources. Auto-classified by AI, deduplicated, confidence-scored, and searchable.

## Agent-Ready Output
Prisma's 9-section dossiers are structured so coding agents (Cursor, Claude Code, GitHub Copilot) can implement them directly. Prisma figures out what to build; your coding agent builds it. The dossier includes Technical Spec and Acceptance Criteria sections designed for implementation.

## Integrations
- Slack (customer feedback, team discussions — OAuth)
- Linear (issues, feature requests, bug reports — OAuth)
- GitHub (issues, pull requests, code architecture analysis — OAuth)
- PostHog (analytics, funnels, feature flags, experiments, errors — API key)
- Notion (documents, wikis, databases — OAuth)
- Jira (issues, epics, sprints, bug reports — OAuth)

## Pricing
14-day free trial with $10 AI credit. Plans: Starter ($20/mo, $20 credit), Pro ($60/mo, $60 credit). All features on every plan.

## Links
- Website: https://tryprisma.ai
- Pricing: https://tryprisma.ai/pricing
- Privacy: https://tryprisma.ai/privacy
- Terms: https://tryprisma.ai/terms
- Download: https://tryprisma.ai (macOS only)
- Full LLM context: https://tryprisma.ai/llms-full.txt