The problem with pasting code into AI chats
When you ask ChatGPT or Claude to help with a bug, a refactor, or a new feature, the quality of the AI's response depends entirely on how much relevant code it can see. Pasting a single file rarely gives enough context. The AI cannot see your types, your helper functions, your configuration, or how modules connect. It fills in the gaps with guesses, and those guesses are often wrong.
The alternative is manually copying multiple files into the chat, which is tedious and error-prone. You forget to include a dependency, or you paste too much irrelevant code and dilute the signal. What you need is a way to assemble exactly the right set of files into a structured bundle that the AI can navigate efficiently.
Riflet vs. CLI tools: Repomix and code2prompt are designed for developers comfortable in a terminal. Riflet is a desktop app with a file browser, preview pane, and token counter. You see what you are including, you check the size, and you export. No installation scripts, no config files, no dependencies.
How developers use Riflet
Open your project folder or connect a GitHub repo
Riflet reads any local folder. On the Pro plan, you can also connect public or private GitHub repositories directly, without cloning them first.
Select the files that matter
Browse the file tree and pick exactly what the AI needs: source files, config, types, tests, README. Deselect node_modules, build artifacts, and anything irrelevant. Preview each file before including it.
Check token count and minify
Riflet shows a token estimate so you can stay within Claude's 200K or ChatGPT's 128K limit. On Pro, content minification reduces token usage by stripping comments and whitespace without changing the code's meaning.
Export and paste into any AI
The output is a single Markdown file with each source file clearly labeled. Paste it into ChatGPT, Claude, Cursor, or any tool that accepts text. The AI sees your code structure, not just isolated snippets.
Combine multiple sources in one bundle
Real development work rarely lives in a single repo. You might need the AI to see your frontend code, your API schema, your deployment config, and a design document. Riflet's Pro plan lets you combine multiple local folders, GitHub repos, and web pages into one bundle, giving the AI a complete picture of your system.
Save the bundle as a workflow and reload it whenever you start a new AI session. When you update your code, refresh the workflow and the bundle picks up the changes automatically.
Works with every AI coding tool
Riflet's output is plain text. It works anywhere: ChatGPT, Claude, Gemini, Cursor, Windsurf, Copilot Chat, local models running in Ollama or LM Studio, or any other tool that accepts pasted text or file uploads. You are never locked into one platform.
Give AI your full codebase context
Download Riflet free and stop pasting isolated code snippets into AI chats.
Download Riflet free