Turn sitemaps into ChatGPT context

Why sitemaps needs preparation before ChatGPT

Large docs sites and marketing sites are hard to collect manually because the important pages are spread across many URLs.

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 a sitemap when you need many related public pages, then filter down to the pages that are actually relevant.

• Add sitemaps as a Riflet source.
• Select docs sections, changelog pages, feature pages, support articles, and public knowledge-base pages.
• 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 docs sections, changelog pages, feature pages, support articles, and public knowledge-base pages. Add a short instruction file at the top if the AI needs naming conventions, project goals, or constraints.

A sitemap can produce too much context quickly. Use filters and token counts to avoid exporting every page just because it exists.

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Project files persist until you remove them, so a stale bundle keeps answering with outdated facts long after the sources moved.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.