The AI-Native Product
Specification Ecosystem
Describe your product in plain English-like specs. Prodara's compiler validates everything, and your AI coding agent turns it into production-ready code - with governance, testing, and incremental builds built in.
$npm install -g @prodara/cliWorks with all major AI providers
26+ AI agents supported
One Ecosystem. Everything You Need.
Prodara isn't just a language - it's a complete toolchain for turning product ideas into production-ready code.
Specification Language
Declarative .prd files that describe entities, workflows, surfaces, governance, and 20+ other constructs - your full product definition in plain text.
Compiler & CLI
15-phase compiler pipeline with type checking, semantic validation, and graph building. Produces deterministic Product Graphs, incremental plans, and impact analysis.
VS Code Extension
Syntax highlighting, real-time diagnostics via LSP, completions, document symbols, hover information, and go-to-definition. The best way to work with Prodara.
AI Agent Workflows
26+ supported agents with 29 slash commands, 5 interactive modes, and structured JSON output. Built to be driven by GitHub Copilot, Claude, Cursor, or any AI coding agent.
How It Works
From idea to production-ready code in three steps.
Describe Your Product
Use the VS Code extension and your AI agent to write .prd specs. Describe entities, workflows, surfaces, governance, and more - or let your agent write it from a prompt.
prodara init --template minimalCompile & Validate
The compiler validates your spec through 15 phases - lexing, parsing, binding, type checking, semantic validation, graph building, and more. Get instant, actionable error messages.
prodara build .Iterate & Ship
Change the spec, rebuild, and see exactly what changed with semantic diffing and impact propagation. 9 built-in reviewers catch issues before they reach code.
prodara plan --format json .First-Class IDE Experience
The Prodara VS Code extension is the easiest way to work with Prodara. Get real-time feedback as you write - or let your AI agent write for you.
- Syntax Highlighting
Rich semantic coloring for .prd files - keywords, types, identifiers, and string literals with a dedicated TextMate grammar.
- Real-Time Diagnostics
See compiler errors and warnings inline as you type, powered by the Prodara Language Server with 300ms debounced validation.
- Smart Completions
Context-aware autocomplete for all PRD keywords, type names, and cross-module references - triggered on "." and ":" characters.
- Agent-Friendly
Your AI coding agent (Copilot, Claude, Cursor, etc.) writes .prd files directly in VS Code with full LSP support, go-to-definition, and hover info.
- One-Click Compile
Build, validate, view the Product Graph, and show the incremental plan - all from VS Code commands.
product task_board {
title: "TaskBoard"
version: "1.0"
modules: [board]
}
module board {
entity task {
task_id: uuid
title: string
status: task_status = todo
assignee: optional
}
enum task_status { todo, in_progress, done }
workflow create_task {
capability: task_management
authorization { user: [task.create] }
writes { task }
returns { ok: task }
}
surface board_view {
kind: view
title: "Task Board"
binds: task
actions: [do_create_task]
}
} Why Prodara?
Whether you're new to specification languages or migrating from another tool, Prodara gives you superpowers.
New to Spec Languages?
Here's why your first spec should be a Prodara spec:
- You don't need to learn the language
Your AI agent (Copilot, Claude, Cursor) writes the .prd specs for you. Focus on what your product does, not syntax.
- Real compiler, real errors
Not a linter on top of markdown - a full lexer/parser/binder/type-checker that catches mistakes before they reach code. 900+ stable error codes.
- VS Code extension included
Install the extension and get syntax highlighting, inline diagnostics, completions, go-to-definition, and hover info from day one.
- Spec-native testing
Write expect blocks right inside your spec. Test transitions, authorization, validation rules, and return types - directly against the Product Graph.
Already Using AI Agents?
Here's what Prodara adds to your workflow:
- Deterministic builds
Same spec always produces the same Product Graph. No hallucinated drift between runs - verified by SHA-256 hashing.
- Compiler-validated specs
A full 15-phase compiler catches mistakes before they reach generated code. Suggested fixes let agents self-correct in a loop.
- Semantic diffing & impact analysis
When you change a spec, Prodara traces 40+ edge types to tell you exactly which downstream artifacts need updating - with depth tracking.
- Governance as code
Constitutions enforce security, privacy, and compliance policies. 9 built-in reviewers plus custom reviewers catch issues automatically.
See It in Action
A complete spec for a to-do app. Write it yourself - or let your AI agent write it for you.
product todo_app {
title: "Todo App"
version: "0.1.0"
modules: [todo]
}
module todo {
actor user { title: "User" }
entity task {
task_id: uuid
title: string
done: boolean = false
created_at: datetime
}
workflow create_task {
capability: task_management
authorization { user: [task.create] }
writes { task }
returns { ok: task; error: task_error }
}
workflow toggle_task {
capability: task_management
reads { task }
writes { task }
}
surface task_list {
kind: view
title: "Tasks"
binds: task
actions: [do_create_task, do_toggle_task]
}
test task_entity_exists {
target: task
expect { kind: "entity" }
}
}Start from a Template
prodara init ships with 5 project templates. Pick one and start building in seconds.
Minimal
Single entity, one file. Perfect for learning Prodara basics.
prodara init --template minimal SaaS
Authentication + billing with Stripe integration patterns.
prodara init --template saas Marketplace
Users, listings, orders - a complete marketplace spec.
prodara init --template marketplace Internal Tool
Admin dashboards and task management for internal teams.
prodara init --template internal-tool API
API keys, resources, and rate limiting patterns.
prodara init --template api Ready to build specs that actually compile?
Install the VS Code extension, point your AI agent at a template, and get your first Product Graph in minutes.