How they fit together
Claude Projects holds your context. Riflet builds the file you put in it. If your sources live in more than one place, a folder, a repository, a documentation site, a stack of PDFs, then assembling them by hand is the slow part, and that is the part Riflet handles.
These are not competing tools. Claude Projects is built by Anthropic and is where your context lives. Riflet is a local-first desktop app that assembles that context from mixed sources and exports it as one portable file. You use both.
Who does what
| Task | Riflet | Claude Projects |
|---|---|---|
| Collecting sources from repos, sites, vaults, and folders | Connects to each source type and pulls the files together in one workspace. | Not part of what it does. It receives files you already have. |
| Choosing what the AI should and should not see | Visual file tree, preview, ignore patterns, and saved filter presets. | You decide at upload time, one file at a time. |
| Reading PDFs, DOCX, XLSX, and PPTX into clean text | Parses them locally into Markdown and fenced code blocks. | Handles common documents directly once uploaded. |
| Knowing the size of the context before you use it | Live token count against the selected model window, before export. | You find out once the knowledge is in place. |
| Keeping context available across conversations | Not its job. Riflet produces the file and stops there. | This is exactly what Project knowledge is for. |
| Using the same context in ChatGPT or Gemini too | The exported file is plain text, so it uploads anywhere. | Project knowledge stays inside Claude. |
What Claude Projects does well
- Project knowledge stays available across every conversation in the project, so you stop re-explaining the same background.
- Custom instructions let you set tone, role, and output format once for the whole project.
- Claude reads the uploaded knowledge directly, with no assembly step on your side.
- It is the natural place to keep long-running work if Claude is your main tool.
What it leaves to you
- It cannot reach out and collect sources for you. A GitHub repository, a documentation sitemap, an Obsidian vault, or a Notion export has to be turned into files first.
- Uploading is manual, and so is keeping it current. When your sources change, you go back and replace what you uploaded.
- You cannot see what the bundle costs in tokens before you commit it, so trimming happens after the fact instead of before.
- Anything you upload lives in Claude. If you also use ChatGPT or Gemini, you prepare and upload the same context again for each one.
Building the file in Riflet
Add your sources in Riflet
Point Riflet at local folders, a public or private GitHub repository, a web page or full sitemap, an Obsidian vault, or a Notion export. Mixed sources in one workspace is the whole point.
Select exactly what belongs in the project
Work through the file tree, preview anything you are unsure about, and use ignore patterns or an imported .gitignore to drop build output, lockfiles, and noise before it reaches Claude.
Check the token count before you export
Riflet shows the live token count for your selection against the model context window, so you find out the bundle is too large while you can still cut it.
Export one file and upload it to Project knowledge
Export a single .txt or .md file, open your Claude Project, and add it to Project knowledge. One file is easier to manage and replace than a scattered pile of uploads.
Build your Project knowledge file with Riflet
Combine files, folders, repos, websites, PDFs, Office docs, and notes into one context file, then upload it where you already work.
Download Riflet freeFrequently asked questions
Do I still need Claude Projects if I use Riflet?
Yes. They do different jobs. Riflet assembles and trims the context file. Claude Projects is where that file lives so Claude can use it across every conversation in the project. Riflet has no chat, no memory, and no conversation history of its own.
Does Riflet sync automatically into Claude Projects?
No. Riflet creates the export and you upload it yourself. When your sources change, you re-export and replace the file in Project knowledge. Saved workspaces make the re-export a one-click step, but the upload stays manual.
Which export format should I upload to Project knowledge?
A single .txt or .md export is the simplest option. Both carry a short preamble explaining the structure, file path separators, and code in fenced blocks with a language tag, so Claude can tell where one file ends and the next begins.
Can I upload the same file to ChatGPT or Gemini?
Yes. That is the practical reason to build the file outside the platform. The export is plain text, so the same bundle works in ChatGPT Projects, Gemini Gems, Perplexity, and local models without being rebuilt for each one.