AI-Powered Project Platform

From Words
to Worlds

Ptah is the AI workspace where ideas turn into specs, working code, and live demos — in hours, not weeks. Built for agencies, startups, product teams, and anyone who ships.

No credit card required
Deploy in minutes

The Problem

The gap between idea and working software is too wide

Whether you're pitching a client, validating a startup idea, or prototyping a feature — the path from "we should build this" to a real, running demo burns weeks of your team's time.

Without Ptah

1

Write specs by hand, iterate with stakeholders

2–3 days

2

Spin up a codebase, wire auth, DB, infra

3–5 days

3

Build the actual feature & UI

1 week+

4

Deploy, fight TLS, share URL

1 day

5

Write estimates & WBS manually

1–2 days

Total time: 2–3 weeks

With Ptah

1

Drop your docs, chat with Claude to shape the spec

Minutes

2

Claude codes the MVP inside a ready-made project

Hours

3

Auto Docker + Nginx + SSL on your VPS

Automatic

4

Live demo URL ready to share

Instant

5

Optional: WBS Excel with hours, roles, costs

Instant

Total time: A few hours

Use Cases

Built for every team that ships

Ptah adapts to how your team works — from winning bids to validating ideas to teaching juniors. One platform, many workflows.

Agencies & Outsourcing

Pain

Dev resources burned on pre-sale demos and WBS spreadsheets.

With Ptah

Win bids with a live demo + costed WBS in the same afternoon.

Startups & Founders

Pain

Weeks lost building throwaway prototypes to test ideas.

With Ptah

Ship a clickable MVP over the weekend — test, pivot, repeat.

Product Teams

Pain

Design → engineering handoff stalls before anything runs.

With Ptah

Turn a PRD into a working prototype the whole team can try.

Consultancies

Pain

Scoping calls end without anything tangible to show.

With Ptah

Walk out of discovery with a proof of concept already live.

Internal R&D

Pain

Engineers want to try new stacks but can't justify the setup time.

With Ptah

Spin up a reference project on any stack, fully deployed, in minutes.

Educators & Mentors

Pain

Students get stuck on setup and never reach the interesting parts.

With Ptah

Give each learner a guided Claude-powered project from day one.

Features

Everything you need to go from brief to live demo

Ptah combines AI analysis, automated development, document collaboration, and project management into one seamless workflow.

AI-Powered Analysis

Claude acts as your BA assistant — analyze briefs, suggest user flows, design module architecture, and expose edge cases automatically.

Auto Code & Deploy

From spec to live demo with Docker + Nginx + SSL. Claude writes the code, builds containers, and deploys to your VPS — fully automated.

Document Annotation Loop

Comment on any PDF, DOCX, or Markdown. Claude reads your feedback, edits the source, and regenerates the output — end to end.

Session Forking

Branch any Claude session with full context inheritance. Explore alternatives, compare approaches, never lose your train of thought.

WBS & Estimates

Auto-generate a Work Breakdown Structure Excel with realistic hours, role assignments, delivery tiers, and cost calculations.

Multi-tenant Workspaces

Organizations, roles, and permissions built for teams. BA, PM, Tech Lead, Founder — everyone gets the view they need.

Under the Hood

Built on deep engineering, not just prompts

Ptah isn't a wrapper around a chat box. It's a full project platform with battle-tested primitives for real software delivery.

Claude Code Sessions

Real Claude Code, scoped per project

Every project gets its own directory, session history, and CLAUDE.md. No context bleed, no lost work between runs.

Annotation Round-Trip

Feedback that closes the loop

Inline comments on PDFs and docs flow back to Claude, which edits the source and regenerates the output — automatically.

MCP Tool Library

19 built-in tools for the agent

Read documents, resolve annotations, register outputs, query sessions — all pre-wired for Claude to use on day one.

Tech Stack Templates

CLAUDE.md rules per stack

Pick a stack, get the right conventions, commands, and guardrails automatically injected into every session.

VPS Auto-Provisioning

From bare SSH to production

One wizard installs Docker, Nginx, Certbot, and the Ptah agent. Wildcard subdomains and SSL handled for you.

Content-Addressed Cache

Attachments by hash, not upload

SHA-256 cache on every VPS means files are deduped across projects and never re-uploaded twice.

How It Works

Four steps from idea to running software

1

Drop Your Docs

Create a project and upload briefs, specs, wireframes, or any reference material — PDF, DOCX, Markdown, anything.

2

Analyze with AI

Chat with Claude to shape requirements, define modules, design user flows, and generate technical specs you can annotate.

3

Build & Deploy

Claude writes the code, containerizes it, and deploys it to your VPS with SSL — you watch it happen in real time.

4

Share & Iterate

A live demo URL, an optional costed WBS, and a workspace where feedback flows straight back into the next iteration.

Integrations

Works with your tech stack

Claude can generate projects with any modern framework or language.

React
Vue.js
Next.js
Nuxt
Angular
NestJS
Spring Boot
FastAPI
Django
Express
Elysia
PostgreSQL
MongoDB
Redis
Docker
Nginx
TypeScript
Python
Java
Go
React
Vue.js
Next.js
Nuxt
Angular
NestJS
Spring Boot
FastAPI
Django
Express
Elysia
PostgreSQL
MongoDB
Redis
Docker
Nginx
TypeScript
Python
Java
Go

AI Providers

Model-agnostic by design

Ptah runs on Claude Code today and the architecture is built so you can bring your own model tomorrow. More providers are on the roadmap — pick the one that fits your budget, latency, and compliance needs.

Claude logo
Available now

Claude

Anthropic

The default agent for every project — fastest path from brief to working demo.

GPT logo
On the roadmap

GPT

OpenAI

GPT-5 and o-series models via the OpenAI API for teams already on that stack.

Gemini logo
On the roadmap

Gemini

Google DeepMind

Gemini 2.5 Pro with long-context support for massive spec documents.

DeepSeek logo
On the roadmap

DeepSeek

DeepSeek

Open-weight reasoning models for teams that need to self-host or trim cost.

Llama logo
On the roadmap

Llama

Meta

Self-hosted Llama variants for air-gapped or on-prem deployments.

Mistral logo
On the roadmap

Mistral

Mistral AI

European-hosted models for GDPR-sensitive engagements.

Qwen logo
On the roadmap

Qwen

Alibaba

Strong coding benchmarks at a fraction of frontier-model cost.

Grok logo
On the roadmap

Grok

xAI

Fresh context + real-time web integration for up-to-date projects.

FAQ

Questions, answered

The essentials before you kick the tires.

You do. Every project lives in a Git repository you control, on a VPS you control. Ptah never locks your code behind a proprietary format.

Yours. Ptah connects to a Claude Code CLI logged in on your VPS, so your usage and billing stay with Anthropic — Ptah is just the orchestration layer.

Any stack Claude can write. We ship first-class templates for React, Next.js, Vue, NestJS, FastAPI, Elysia, Django, Spring Boot, and more — and you can add your own.

Yes. Ptah Cloud manages the workspace, but code and demos run on VPS workers you own. One install script sets everything up.

Documents live in your own S3 bucket. Chat sessions stream directly between your VPS and your browser. Ptah never trains on your content.

No — that was the original pain point, but teams also use Ptah for POCs, internal R&D, product prototyping, consulting engagements, and teaching.

Ship your next idea this week

Stop burning weeks on demos and proposals. Start turning briefs into running software today.

Get Started Free