ShiftStack FAQ
Common questions about ShiftStack, AI Builder, App Doctor, context attachments, and where the platform is headed.
General
What is ShiftStack?
ShiftStack is a project management and AI-assisted builder tool for web studios and freelancers. It helps you track client projects, generate starter code using AI, and diagnose common app issues — all from one dashboard. It is being built toward a broader multi-platform AI builder workflow, rolling out features in stages.
How is ShiftStack different from Vercel, Bolt, Base44, Webio, or Manus?
ShiftStack is not yet at the same level as those tools. Vercel is a world-class deployment platform. Bolt, Base44, Webio, and Manus are more advanced AI builder products. ShiftStack is purpose-built for web studios managing client projects, with AI code generation and diagnostics layered on top. We are honest about where we are today and what is planned for future stages.
Does ShiftStack execute or deploy code automatically?
No. ShiftStack never runs, executes, installs, or deploys your code at any stage. The AI Builder generates copy-ready source files, and the optional GitHub export commits those files to a new repository — but nothing is run or deployed automatically. You always review and deploy on your own terms.
Is ShiftStack free to use?
ShiftStack is currently in active development. Pricing details will be announced as the platform matures. The current version is available to early users while features are rolled out.
AI Builder
Can I attach a PRD or existing code to AI Builder?
Yes. In both Build and Diagnose modes, you can attach up to 5 files (PRD, code, schema, log, etc.) so the AI has real context to work from. Text is extracted client-side from each file and sent to the AI as part of your prompt. PDF and DOCX files are parsed in the browser — no server upload is needed. Persistent attachment storage across sessions is planned for a future stage.
What file types can I attach?
Readable text formats: .txt, .md, .json, .csv, .html, .css, .js, .jsx, .ts, .tsx, .sql, .yml, .yaml, .xml, .log, .env.example, .pdf, .docx. PDF and DOCX files are parsed client-side and their extracted text is sent as context. If a PDF or DOCX cannot be parsed (e.g. scanned image-only documents), you will see an error suggesting you export as TXT or MD instead. Images (.png, .jpg, .jpeg, .webp) and ZIP files are planned for future support. Limits: 1 MB per file, 3 MB total combined, 5 files maximum per submission.
Does ShiftStack execute uploaded code?
No. Attached files are treated as read-only context for the AI — their text content is included in the prompt so the AI can reference your real code or schema. Nothing is executed, imported, installed, or run at any point. This is the same guarantee that applies to generated code.
Should I upload API keys or secrets?
No. Never upload files containing real API keys, passwords, database credentials, or payment data. If you want to share environment variable structure, use a .env.example file with placeholder values (e.g. ANTHROPIC_API_KEY=your-key-here) instead of real values. ShiftStack instructs the AI to never reproduce secrets it sees in context.
Can App Doctor use attachments to debug issues?
Yes. When you run a diagnosis, any attached text files — error logs, stack traces, schema files, relevant code snippets — are passed to App Doctor as context. The diagnosis will reference the files it used, shown under “Context used” in the results panel.
What is the difference between Build mode and Diagnose mode?
Build mode generates copy-ready source files based on a prompt you write. You pick a platform and describe what you want built. Diagnose mode (App Doctor) takes an issue description and returns a plain-English diagnosis with likely causes, checks to perform, and a recommended fix plan — no code is generated or modified automatically.
What platforms does AI Builder support?
Currently active: Next.js, React, HTML/CSS/JS, Supabase, Node.js API, and Existing app (generic). Shopify and WooCommerce are listed as planned targets and will be supported in a future stage as those integrations are built out.
Can I push generated code to GitHub?
Yes, via Stage 2. After code is generated, you can create a new GitHub repository and commit all generated files to it. This requires a GITHUB_TOKEN environment variable with repo scope. The files are committed to the new repository — nothing is installed or deployed.
Projects
What project statuses are available?
Draft, In Progress, and Delivered. You can filter and search projects by status from the dashboard overview.
Can I export my project data?
Yes. Use the Export button on the dashboard overview to download all your projects as a structured JSON file. This is useful for backups, migrations, or sharing project data with other tools.
Technical
How is my data stored?
Project data is stored in Supabase with Row Level Security (RLS) enabled. Your data is only accessible by your own authenticated account. Supabase handles authentication, database, and session management.
What AI model does ShiftStack use?
ShiftStack uses Anthropic Claude for all AI features — streaming chat, code generation, and App Doctor diagnostics. This is configured server-side via ANTHROPIC_API_KEY — the key is never exposed to the browser.
Does ShiftStack store my prompts or generated code?
Currently, prompts and generated results are not persisted to the database. They exist only in your browser session. Saving generation history is planned for a future stage.
Still have questions?
Jump into the dashboard and try AI Builder or App Doctor directly — no setup required beyond signing in.
Go to Dashboard