The platform lock-in problem
Both OpenAI and Anthropic have built features for persistent project context. ChatGPT has Custom GPTs and Projects. Claude has Claude Projects. Both let you upload files that persist across conversations within their platform.
But these systems are closed. A Claude Project cannot be opened in ChatGPT. A Custom GPT cannot be loaded in Claude. If you use both tools (and many people do, because different models excel at different tasks), you end up maintaining separate, duplicate sets of project files on each platform. When your project evolves, you have to update both. Context drifts. Results become inconsistent.
The portable alternative: Build a single context bundle that lives on your computer. Paste or upload it into whichever AI tool you are using that day. The same file works in ChatGPT, Claude, Gemini, Perplexity, local models, and any other tool that accepts text.
How a portable context bundle works
The concept is straightforward. Instead of uploading your project files into a platform-specific system, you merge them into one structured text file on your computer. This file contains your project brief, reference documents, style guidance, and instructions for the AI. It is a standard Markdown or plain text file, not tied to any platform.
When you start a session in ChatGPT, you paste or upload the file. When you switch to Claude, you paste or upload the same file. The AI reads it, gets full project context, and you start working. There is nothing platform-specific about a text file.
Step-by-step workflow
Assemble your project files locally
Gather the files that define your project: notes, documents, code, research, specifications. These live on your computer, not inside any AI platform.
Build one bundle with Riflet
Open Riflet, select your files, and export a single structured Markdown file. Save it as a workflow so you can reload it in one click. The bundle is a local file under your control.
Use it everywhere
Paste the bundle into ChatGPT when you want GPT-4o's reasoning. Paste it into Claude when you want Claude's long-form writing. Paste it into Gemini when you want a third perspective. The AI sees the same project context every time, regardless of platform.
Update once, use everywhere
When your project changes, update the bundle in Riflet. The new version is automatically available for every AI tool. No need to update separate Claude Projects and ChatGPT configurations independently.
Why this approach is better for multi-model users
Different AI models have different strengths. Claude excels at long-form analysis and careful reasoning. ChatGPT is fast and broadly capable. Gemini has strong integration with Google's ecosystem. Using multiple models for different tasks is a practical strategy, not an edge case.
A portable context bundle lets you leverage each model's strengths without duplicating your project setup. You maintain one source of truth for your project context, and every AI sees the same information. This eliminates the consistency problems that come from maintaining separate project configurations on different platforms.
Works with local AI models too
If you run local models through Ollama, LM Studio, or similar tools, the same bundle works there as well. Paste the text into the prompt, and your local model gets the same project context as the cloud models. This is particularly valuable for users who need to keep sensitive data local while still using AI assistance.
One bundle, every AI platform
Build your portable context bundle with Riflet. Download free for macOS, Windows, and Linux.
Download Riflet free