PromptShift is a sovereign developer tool that translates entire AI assistant contexts, instruction files, master prompts, and system rules between Cursor (.cursorrules / .mdc), Anthropic Claude Projects, OpenAI Custom GPTs, and Google Gemini / Antigravity in one click.

Is my source code and proprietary API keys secure?

Yes. PromptShift operates with strict in-RAM processing and zero server retention. All uploads are processed ephemerally and sanitized in memory to redact leaked API keys (OpenAI, Anthropic, Stripe, AWS, HuggingFace, etc.) before conversion. Nothing is saved to a persistent server disk or database.

What are the pricing options and tiers?

PromptShift offers a Free Starter tier (instant conversion for projects up to 2 files with zero signup), a Single Migration Pass at $1.99 for one-off full project exports, a Pro Developer plan at $5.99/mo for unlimited multi-ecosystem migrations, and a Team & Agency pass at $19.99/mo.

Which AI platforms are currently supported?

PromptShift supports bidirectional migration between Anthropic Claude (.claude / system instructions), Cursor (.cursorrules / .cursor/rules/*.mdc), OpenAI Custom GPTs / Assistants v2, Google Gemini & Antigravity, and GitHub Copilot Workspace.

Can I convert back and forth without losing rules?

Yes! PromptShift uses a unified AST intermediate representation (CAPS) that preserves 98.5%+ certified semantic fidelity. Autonomy directives, tone, architecture policies, and negative constraints are mapped without prompt drift.

How do I manage my billing, receipts, and VAT invoices?

All checkouts and subscriptions are securely processed by Polar.sh as Merchant of Record. You can access your 1-click self-service customer portal directly inside PromptShift to manage billing, download tax-compliant VAT invoices, and adjust plans anytime.

98.5% Certified Semantic Parity•100% In-RAM Zero-Retention

Convert GitHub Copilot Instructions to Cursor Rules

Split monolithic .github/copilot-instructions.md into modular, high-performance Cursor rules (.cursorrules and .cursor/rules/*.mdc).

🚀 Drop & Convert (GitHub Copilot → Cursor)⚡ Unlock 4-in-1 Suite ($1.99)

01Architecture & File Mapping

Source Project Structure

GitHub Copilot

📄.github/copilot-instructions.md
📄.github/instructions/*.instructions.md

Transpiled Target Structure

Cursor

✨.cursorrules
✨.cursor/rules/*.mdc
✨.cursor/settings.json

GitHub Copilot Directive	Cursor Equivalent	Semantic Bridge Behavior
.github/copilot-instructions.md	.cursorrules	Extracts core global architectural directives for Cursor
Copilot section headings	.cursor/rules/*.mdc	Splits frontend, backend, test, and database rules into glob-scoped .mdc files

02Step-by-Step Migration Walkthrough

Upload your GitHub Copilot workspace files

Drag and drop your project repository or specific rule files (.github/copilot-instructions.md, .github/instructions/*.instructions.md) into the PromptShift workspace. All reading and parsing is performed strictly in your browser RAM without disk persistence.

Inspect Automated Secret Sanitization & AST Fidelity

PromptShift executes active secret sanitization in memory, redacting exposed API keys (OpenAI, Anthropic, Stripe, AWS, HuggingFace) before compiling to the Canonical AI Project Schema (CAPS) with certified 98.5% semantic parity.

Export Ready-to-Run Cursor Package

Click download to receive your fully structured ZIP archive containing .cursorrules, .cursor/rules/*.mdc, .cursor/settings.json. Drop these files directly into your workspace to instantly begin using your AI assistant.

03Why Developers Choose PromptShift for GitHub Copilot → Cursor

✓

Parses monolithic Copilot instructions into scoped .cursor/rules/*.mdc modules

✓

Automatically generates valid YAML frontmatter with quoted globs

✓

Optimizes context window by preventing unnecessary rule injection

✓

In-RAM secret scrubber eliminates accidentally committed tokens

04Frequently Asked Questions

Why convert Copilot instructions to Cursor .mdc rules?

Monolithic copilot-instructions.md files load into every prompt turn. PromptShift splits them into targeted .mdc rules that only load when relevant files are edited, cutting token waste by ~50%.