PromptArch v2 Is Here
Today we're releasing PromptArch v2, the biggest update since launch. The headline: a brand-new Context Engineering Studio that lets you build, score, and optimize the text artifacts that power modern AI systems.
If you've been hand-writing system prompts, cursor rules, or CLAUDE.md files with no structured guidance, this is for you.
Why Context Engineering?
The field has shifted. In 2024, "prompt engineering" meant finding the right words for a single instruction. In 2026, the best AI practitioners are designing entire information environments: system prompts, tool definitions, agent configurations, memory files, and structured metadata that together shape how AI behaves.
Gartner calls it context engineering. Anthropic, OpenAI, and every major AI lab now builds around this paradigm. But here's the gap: everyone is building the infrastructure for context engineering (MCP servers, agent frameworks, orchestration layers), while nobody is building tools to help people write the text artifacts that go into those contexts.
That's what PromptArch v2 is for.
Read our deep dive: Context Engineering: Why Prompt Engineering Is No Longer Enough.
The Context Engineering Studio
The Studio is a structured, guided editor for the 12 artifact types that define how AI models behave across tools and platforms:
IDE & Coding Agents
- Cursor Rule:
.cursorrulesand.mdcfiles that tell Cursor how to write code in your repo - Windsurf Rule: Configuration files for Windsurf IDE
- Copilot Instructions: GitHub Copilot custom instructions
- CLAUDE.md: Repository-level context files that Claude Code reads on every interaction
- AGENTS.md: Multi-agent coordination files for complex coding projects
Claude Ecosystem
- Claude Code Skill: Reusable SKILL.md files for Claude Code
- Claude Project: Instructions for Claude Projects (long-running conversations with custom context)
ChatGPT & GPT
- Custom GPT: System prompts for GPT Store applications
- GPT System Prompt: API-level system prompts for GPT-powered applications
Cross-Platform
- MCP Tool Description: Tool descriptions and schemas for Model Context Protocol servers
- Agent Task Spec: Task specifications for multi-agent orchestration (LangGraph, CrewAI, AutoGen)
- System Prompt (General): Model-agnostic system prompts for any LLM application
Every artifact gets PromptArch's signature treatment: a guided step-by-step wizard, model-specific optimization, and a 4-dimension quality score with actionable suggestions.
What Else Is New in v2
- Updated positioning: The homepage and hero section now reflect PromptArch's expanded scope: from prompt engineering to context engineering
- Research page: New section on context engineering with academic backing
- Context engineering landing page: A dedicated page explaining the concept and how PromptArch maps to it at /context-engineering
Coming from PromptPerfect?
If you're one of the users affected by PromptPerfect's shutdown, PromptArch is built to be a stronger alternative. We offer domain-specialized guided wizards (not just a text box optimizer), 4-dimension quality scoring, model-specific optimization, and flexible pricing: subscribe for the best rate or just buy credits that never expire.
Check out our migration guide for a step-by-step walkthrough.
Try It Now
The Context Engineering Studio is available today. Head to /studio to start building the text artifacts that power your AI.
Every new account gets free credits to get started, no credit card required.