ChatGPT searches your long PDFs instead of reading them — and never tells you
Ok_Negotiation_2587 · reddit · 2026-09-22
A Reddit PSA based on real testing: ChatGPT fully reads short files, but long documents (contracts, theses, 200-page reports) mostly go into a search index, and answers are built from whatever chunks match your question. The interface never indicates which mode is in play, and answers sound equally confident either way.
The author's cautionary tale: a clean summary of a long agreement missed the one clause they actually needed near the end, because "summarize this" didn't point the search at it; asking about the clause by name found it instantly.
Workarounds:
- Ask about sections by name or page ("What does section 9 say about termination") instead of whole-file prompts.
- For real summaries, go section by section ("summarize pages 1–20") so every part is actually read.
- Ask which sections it didn't use — it's often honest.
- Split long files before uploading; several shorter files get read far more completely.
The search behavior is native to ChatGPT, unrelated to the author's extension.
More from Apps
- Ditto launches swap-anything video tool to replace any video's main character — yoobinray · 2026-11-12
- Loom founder launches Roma 2.0, a shared to-do list that orchestrates AI agents — thetripathi58 · 2026-11-12
- AI pipeline turns a few product photos into a full video ad via storyboard-first workflow — jeffgrimes9 · 2026-11-12
- Pika Integrates Seedream 5.5: Split Images Into 100+ Layers for Precise Multi-Turn Edits — aziz4ai · 2026-11-12
- Llama app adds simultaneous voice, image and video input, running locally on LFM3-Omni — mervenoyann · 2026-11-12
- Google Labs launches Tabletop, an AI experiment that builds custom shared spaces on demand — GoogleLabs · 2026-11-12