--- title: "Trace what ChatGPT forgot and store it where it stays" description: "ChatGPT forgets what you told it because each fact lands in a different place, so trace where it went and move business facts to a file." canonical: "https://scalewithsearch.com/articles/chatgpt-keeps-forgetting-what-i-told-it" date: "2026-01-27" modified: "2026-09-25" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # Trace what ChatGPT forgot and store it where it stays. You explained your business to ChatGPT three times this week: the industry, the target customers, the brand voice. Each new conversation acted as if you had never met. You told it once to remember your pricing rule, and a week later it quoted the old price. This is not a fault you can report. It follows from how ChatGPT stores what you tell it. A fact you type can land in one of several places, and each place keeps it for a different length of time. Find the place, and you know why the fact slipped. ## Know why a new chat starts without you Each ChatGPT conversation runs inside its own context window: the text the model can read at one time. When you open a new chat, the model does not receive the full text of the old one. Unless a feature carries a fact across, or you paste it in again, the fact is gone for that chat. OpenAI added memory features to carry some facts across. They store fragments, such as your name, your job title, or a preferred reply length. They do not store the operating detail that makes an output useful: client profiles, service terms, pricing rules, and the reasons behind past decisions. The saved-memory store also has a capacity. OpenAI does not publish a fixed count of saved facts, and ChatGPT has shown a "Memory full" notice when the store reached its limit. A single client contract holds more context than that store can carry. ## Find where the fact went A fact that you gave ChatGPT sits in one of six places. Each place has its own rule for when the fact comes back. | Where the fact went | How it got there | When it comes back | |---|---|---| | The chat only | You typed it in a normal message | Only in that chat, while it stays inside the context window | | Saved memory | ChatGPT saved it, or you said "remember this" | When the model judges it relevant; not in every chat | | Chat history reference | ChatGPT drew on a past chat | Selectively; it does not keep every detail | | Custom instructions | You typed it in the settings field | In every ordinary chat, as guidance the model may weigh against your prompt | | A Project | You added it to Project instructions or files | Inside that Project only | | A temporary chat | You used temporary mode | Never; the mode does not use or create memory | Most "it forgot" cases fall in the first two rows. You stated the fact in a message and assumed it was saved. Or it was saved, and the model did not judge it relevant to the new request. The [guide to ChatGPT memory that does not apply](/articles/chatgpt-memory-not-applied-every-conversation) tests the second case across five chats. ## Test the fact in four steps Use a harmless fact for the test, such as a preferred sign-off. Do not test with client data. 1. Open ChatGPT settings and find the memory controls under Personalization. Note whether saved memory and chat history reference are on. Labels change between releases, so record the date and plan. 2. Open the list of saved memories and search for the fact. If it is absent, it was never saved; it lived in a chat only. 3. Open a new chat, not a Project and not a temporary chat. Ask a question whose answer needs the fact. 4. Record whether the answer used the fact, ignored it, or used an older version. If the fact is saved and still does not apply, the model's relevance choice is the cause. If the fact is not saved, the chat was the only place it lived. Both results point to the same conclusion: a fact that must apply every time needs a place that is not selective. ## Count the cost of repetition Ten minutes of explanation before you can ask the real question, five days a week, comes to about 50 minutes a week. Over a working year, that is about 40 hours spent bringing a tool up to speed. The quality cost is larger. When you shorten your context to save time, the output turns generic. When you leave details out, the model fills the gaps with assumptions. Some assumptions are wrong, and you spend the saved time on edits instead. It is like an assistant who arrives each morning with no memory of the day before. You spend the first hour on a briefing. Then the assistant leaves, forgets everything overnight, and you repeat the briefing tomorrow. ## Know why custom instructions do not close the gap The custom instructions field is the first fix most owners try. In January 2026, it held 1,500 characters, about 250 words. As checked on September 19, 2026, OpenAI lists 1,500 characters for Free and Go accounts and 5,000 for paid plans. Even 5,000 characters forces a choice. Your business model, client profiles, services, brand voice, industry terms, and preferred formats do not all fit. So you keep fragments. The fragments carry a simple task, but a complex task still needs a manual explanation. The field is also guidance, not a rule. When your prompt conflicts with it, the prompt usually wins. The limit underneath is structural: the model keeps only what fits in its active window, plus what a memory feature chooses to bring. The [explanation of why AI does not remember context](/articles/ai-doesnt-remember-context) covers that mechanism. ## Move business facts into a record that loads every time The fix is not a better prompt. It is also not a rotation through chats, Projects, and custom GPTs in the hope that one of them sticks. The fix moves your context out of the model's features and into a structured record that loads at the start of each session. The record holds your business details, processes, client information, and preferences. It sits on storage that you control, and an agent reads it without a request from you. The text that the agent reads still goes to the model provider for processing. Some agents support this directly. Claude Code reads a `CLAUDE.md` file from the project folder at the start of every session. If the folder is an Obsidian vault, the file can point to the rest of your notes. The record then survives new sessions, product updates, and a change of model. In ChatGPT, attach the same file to a Project or paste it into an important chat. That step stays manual, but the file remains the single source. You edit a fact once, not in each chat that repeated it. Sort what you told ChatGPT into two groups. Personal preferences, such as reply length or a sign-off, can stay in saved memory or custom instructions. Business facts, such as prices, client terms, and approval rules, move into the file with an owner and a date. The [guide to making ChatGPT remember previous conversations](/articles/make-chatgpt-remember-previous-conversations) shows how to promote facts out of old chats into that record. ## End the workarounds Most advice for this problem asks you to adapt: write better prompts, keep conversation starters, build templates. That advice treats the limit as permanent and leaves the repetition with you. The repetition ends when the context lives in a file instead of in your head. The [broader guide to ChatGPT forgetting your business](/articles/chatgpt-forgets-everything) walks through a starter file and a one-sitting setup. ## Related: AI memory - [ChatGPT Memory Is Wrong: Correct the Record and Test the Next Conversation](/articles/correct-wrong-chatgpt-memory) - [How to Make ChatGPT Remember Previous Conversations Without Treating History as Truth](/articles/make-chatgpt-remember-previous-conversations) - [Move your business context into a file so you explain it once](/articles/chatgpt-forgets-everything) ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```