Can I share project rules between different AI assistants?

With Vilix you write your project conventions and personal style directives once, and every AI assistant you connect reads them automatically before each reply. No per-app custom instructions to maintain.

Yes, you can share project rules between different AI assistants, and Vilix is built specifically for this. You define your rules once inside Vilix, and every connected tool reads them automatically before replying. There is nothing to copy-paste, no per-app custom instruction panels to maintain, and no risk of one tool drifting out of sync with another.

The problem: every AI assistant is an island

Every major AI tool ships with its own custom instructions or system prompt field. ChatGPT has “Custom Instructions.” Claude has Projects with custom instructions. Cursor has a .cursorrules file. Codex has a system prompt. If you want all of them to know that your project uses TypeScript strict mode, that you prefer short replies, or that the API lives at a specific base URL, you end up copying the same block of text into every app and keeping each copy in sync forever.

That is not a workflow. It is a maintenance burden that grows every time you start a new project or add a new tool to your stack.

How Vilix solves it with user_rules and project_rules

Vilix has two rule layers that live server-side in your account and travel with you across every tool you connect.

user_rules: personal style directives

user_rules are short personal directives that apply everywhere, regardless of which project you are in. Examples: “keep answers under three paragraphs,” “never suggest installing a new dependency without asking first,” or “use British English spelling.” You set them once. Every AI that connects to Vilix receives them on every turn, automatically.

project_rules: per-project stack and conventions

project_rules attach to a specific Vilix project and carry technical context: the framework, the conventions, the tone, the things the AI should always or never do for this codebase. A typical entry might look like: “This project is a Next.js 14 app with TypeScript strict mode. Always use the App Router. Never use default exports for components. Tailwind only, no inline styles.”

When you open that project in Vilix and then switch to ChatGPT, Claude, Cursor, or Codex, the AI auto-loads those rules before it replies. You do not mention them in the chat. They just arrive.

How the auto-injection works

Vilix is an MCP-native memory layer. You connect it once as a custom MCP server at api.vilix.ai/mcp. After that, the AI calls two Vilix tools automatically on every turn.

• get_context runs before the AI replies. It pulls your recent conversation history, saved memories, semantically related past exchanges, user_rules, and the project_rules for whichever project is active.
• save_turn runs after the AI replies to persist the exchange so future turns and other tools can see it.

You do not orchestrate this. The AI handles it. The result is that every reply, in every tool, starts with the same rules already loaded.

A concrete example: sharing rules between ChatGPT and Claude

Say you are building a SaaS product. In Vilix you create a project called “Billing service” and add project_rules:

• Stack: Node.js, Stripe, PostgreSQL
• Always wrap Stripe calls in try/catch and log the error to Sentry
• API routes follow REST conventions, no GraphQL
• Short answers preferred; include code snippets when the question is technical

You also have a user_rule: “Never start a reply with the word ‘Certainly.’”

Now you ask ChatGPT to write a new webhook handler. It reads the project_rules and user_rule via Vilix before answering. Later you switch to Claude to review the same handler. Claude reads the same rules from the same Vilix account. You get consistent, convention-aware output from both tools without repeating yourself.

What tools are supported

Any tool that supports a custom MCP connector works. That includes ChatGPT, Claude, Claude Code, Cursor, Codex, Grok, Manus, GitHub Copilot, Windsurf, Lovable, OpenClaw, and Hermes. If a tool can point to an MCP endpoint, it connects to Vilix and inherits your rules.

For a deeper look at what cross-tool memory makes possible beyond just rules, see the cross-AI memory overview.

Setting up your first rules

After connecting Vilix to your first tool, you can add rules in a few ways. You can tell the AI directly: “Add a user rule: always prefer async/await over promise chains.” The AI writes it to your Vilix account on the spot. Or you can manage rules through the Vilix features panel directly.

Rules are versioned in your account, not scattered across app settings. When you refine a rule, every connected tool picks it up on the next turn. No syncing required.

If you want the same conventions everywhere without copying anything, the fastest path is to try Vilix free and add your first project_rules in the first session. From that point forward, every AI you connect inherits them.

Frequently asked questions

Can I share project rules between ChatGPT and Claude?

Yes. Connect both tools to Vilix as a custom MCP server at api.vilix.ai/mcp. Both then read your project_rules automatically on every turn, so they operate from the same conventions without any copy-pasting.

What is the difference between user_rules and project_rules in Vilix?

user_rules are personal style directives that apply across every project and every tool: things like preferred reply length, tone, or habits you want the AI to avoid. project_rules attach to a specific Vilix project and carry technical context: the stack, the conventions, the dos and don’ts for that particular codebase.

Do I need to mention my rules in every chat message?

No. Vilix injects them automatically via the get_context tool call that runs before each AI reply. You write the rules once; the AI receives them silently on every turn without you having to paste or reference them.

What happens if I update a rule? Do all tools pick it up?

Yes. Rules live server-side in your Vilix account. The next time any connected tool calls get_context, it receives the updated version. There is no per-app update step.

Which AI tools support Vilix rules?

Any tool that accepts a custom MCP connector: ChatGPT, Claude, Claude Code, Cursor, Codex, Grok, Manus, GitHub Copilot, Windsurf, Lovable, OpenClaw, and Hermes. If the tool can point to an MCP endpoint, it connects and inherits your rules automatically.