Turn local files into ChatGPT context

Why local files needs preparation before ChatGPT

Real project context often lives in a messy local folder with PDFs, Word docs, spreadsheets, notes, and screenshots next to each other.

ChatGPT is useful for general analysis, writing, coding help, and iterative work when the context is packaged into a readable file. The result depends heavily on how cleanly your source material is grouped, labeled, and filtered before it reaches the model.

Step-by-step workflow

Add the folder that contains your project documents, then use the file tree to keep only the files that support the task.

• Add local folders as a Riflet source.
• Select PDFs, DOCX files, spreadsheets, Markdown notes, plain text files, briefs, and source documents.
• Use filters to remove stale exports, generated files, and unrelated background material.
• Export one Markdown or text bundle and use it in ChatGPT Projects.

What to include

For this workflow, prioritize PDFs, DOCX files, spreadsheets, Markdown notes, plain text files, briefs, and source documents. Add a short instruction file at the top if the AI needs naming conventions, project goals, or constraints.

Binary media should be described or converted before it becomes useful LLM context. Text-heavy documents usually carry the most value.

Getting the bundle into ChatGPT Projects

ChatGPT Projects keeps uploaded material in project files. Create a project and attach the bundle to it, so every chat started inside that project can read the file without you pasting anything. It accepts text, Markdown, PDF, and Office documents, so a Markdown or plain-text export from Riflet needs no further conversion.

ChatGPT retrieves passages from attached files rather than reading each one end to end. Explicit file-path headers and section markers matter more here than anywhere else: a bundle with clear boundaries gets retrieved reliably, an undifferentiated wall of text does not.

Project files persist until you remove them, so a stale bundle keeps answering with outdated facts long after the sources moved. Replace it on the same schedule as the work.

Token budget advice

Riflet lets you check the bundle against GPT-6 Astra's 1,050,000-token context window before export. Keep room for your actual question and for the model's answer.

If the bundle is too large, remove files that merely repeat what another file already explains. Context quality usually beats context volume.

Honest limitation

Riflet does not push live updates into ChatGPT. It creates the context file. When the source changes, re-export and upload the new bundle yourself.