Free tools Windows power users keep installed
One-click scans. No signup required.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
A coding agent mentioning an old coding pattern can feel like being called out. That does not mean it formed a human-like judgment about you—or that every agent keeps a permanent record of every mistake. The detail may have come from chat history, a saved preference, repository notes, project instructions, or files and tools in the current workspace. Those sources have different scopes and controls.
What “remembering” can mean in a coding agent
There is no single, universal memory feature. An agent may build its current context from the conversation, workspace files, tool results, custom instructions, explicit references, or persistent memory, depending on the product. Microsoft says Visual Studio Code agents can use conversation history, workspace files, tool outputs, custom instructions, and explicit references; some of those are limited to a chat, while others can persist. VS Code’s documentation explains its context and memory model.
That distinction matters when a detail seems personal. A code convention found in a project file is not the same thing as an account-level preference, and neither proves that the model kept a permanent log of your errors. The wording “judged me” describes how the response felt; it is not evidence that the agent made a human-like judgment.
Three scopes in Visual Studio Code
VS Code documents three memory scopes, stored locally: user memory for preferences across workspaces, repository memory associated with the current workspace, and session memory for the current conversation. A repository-level note might explain a recurring code pattern; a user-level note might capture a preference; session memory can help maintain continuity within a conversation. These scopes should not be assumed to work the same way in other products. Microsoft describes the scopes and their behavior.
#1 Best Overall
- Careercup, Easy To Read
- Condition : Good
- Compact for travelling
Repository facts and preferences in GitHub Copilot
GitHub describes Copilot Memory as holding repository-level facts and user-level preferences. Repository facts can include coding conventions, architecture decisions, build commands, and project-specific rules. GitHub says those facts carry citations to supporting code and are checked against the current branch before use; only validated facts are used. GitHub also says unused facts or preferences are automatically deleted after 28 days, though successful validation and use may reset the timer. That is a Copilot-specific rule, not an industry-wide retention standard. See GitHub’s Copilot Memory documentation.
How to find the source of a surprising reference
Start by asking the agent what context it used, then check the relevant product controls and project sources. This is a practical troubleshooting sequence, not a guarantee that every product can trace an answer to a particular source.
- Ask what informed the answer. Request the specific context or memory behind the reference. If the product exposes citations or memory entries, inspect them rather than assuming the source.
- Check the active workspace. Look for project instructions, memory files, relevant source code, and other files the agent could read. A team convention may be documented locally rather than stored in your account.
- Review memory and past-chat controls. Check the product’s settings for saved memories, past-chat search, or chat history. Names and effects differ by product.
- Correct or remove the right source. If the detail is wrong or unwanted, edit or delete the memory where possible. If it came from a chat, file, or connected source, address that source separately.
- Move stable team guidance into the repository. Once a convention is verified, document it in a reviewed, source-controlled file or project instructions so the team can see and maintain it.
Memory controls differ by product
| Product | Documented controls or behavior | Important scope or deletion detail |
|---|---|---|
| ChatGPT | Memory controls are under Settings > Personalization > Memory. Depending on plan, region, platform, and workspace, users may be able to review or correct memory, ask ChatGPT not to mention something, or delete a saved memory. OpenAI documents ChatGPT memory controls. | Turning Memory off does not delete past chats. Deleting a chat alone may not remove a separately saved memory. Removing remembered information may require deleting the memory and its original chat, as well as relevant sources such as Library files or connected apps. |
| Claude | Anthropic documents settings to view and edit memory, turn off memory or past-chat search, and use incognito chats. See Claude’s chat search and memory help page. | Anthropic says incognito chats are not saved to memory or chat history. Do not assume these controls or their effects match another product’s. |
| GitHub Copilot | Users can review and delete user-level preferences; repository owners can review and manually delete repository facts. Enterprise and organization administrators have additional management options. GitHub explains Copilot Memory management. | Copilot Memory is enabled per user, with policy behavior that differs across individual and managed plans. Check current account and organization settings rather than assuming one default applies to everyone. |
| Visual Studio Code | VS Code describes user, repository, and session memory scopes, stored locally. Microsoft’s memory documentation details these scopes. | Move verified, stable team guidance into source-controlled documentation or custom instructions and review the change through the normal repository process. |
Deleting a memory may not delete its source
In ChatGPT, a saved memory and the conversation or other source that supplied it are separate. OpenAI cautions that turning off Memory does not delete past chats, and deleting a chat does not necessarily delete a saved memory. To remove remembered information, you may need to delete both the saved memory and the original chat, and check other sources such as Library files or connected apps. OpenAI lists the relevant memory and deletion controls.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →For other products, use that product’s own documentation and settings: a memory control, chat-history control, local project file, and organization policy are not interchangeable. Product behavior and availability can change, and settings may vary by plan, region, platform, or workspace.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Make team conventions explicit and reviewable
Local or agent-managed memory can be useful for continuity, but important team rules should not depend on a hidden or personal note. VS Code advises moving verified, stable guidance into source-controlled project documentation or custom instructions. That makes the convention visible to collaborators, reviewable like other repository changes, and easier to correct when the codebase changes. It also separates shared project knowledge from individual preferences.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
