llm-kb
← Back to news
Incident

Users report ChatGPT Saved Memory bug after plugin use

A reproducible bug in ChatGPT causes Saved Memory to stop functioning after a plugin is used in the same conversation. The model claims to save new memories, but they do not appear in the Manage Memory settings. This affects user workflows relying on persistent context.

3 engagement·1 source·Mon, Jul 13, 2026, 10:17 AM
A Reddit user reported a systemic bug in ChatGPT's Saved Memory feature. When a user asks ChatGPT to remember a test phrase before using any plugin, the memory saves correctly and appears under Settings → Personalization → Manage Memory. However, after using a plugin in the same conversation, subsequent memory saves are not persisted despite ChatGPT confirming the save. The user provided reproduction steps and noted the bug is reproducible. This impacts developers and power users who depend on memory for context across plugin interactions.

Entities

OpenAI(company)ChatGPT(tool)Saved Memory(concept)

Related

Incident

Users report ChatGPT project isolation and memory leaks

A Reddit user reports that ChatGPT's project isolation feature is leaky, causing the AI to bring up the user's project 'Praana' in unrelated chats. The user contrasts this with Claude, which handles context more consistently. The complaint highlights a UX engineering gap at OpenAI.

5 engagement·1 source·reddit
Sun, Jul 12, 2026, 04:31 PM
Incident

Users report missing chat history after installing new Mac ChatGPT desktop app

A user reports that after downloading the new Mac ChatGPT desktop app, their existing projects and past chats disappeared, and the app added unwanted work functions like Codex and Slack integration. The user reverted to the browser version, expressing frustration.

13 engagement·1 source·reddit
Mon, Jul 13, 2026, 09:44 AM
Product

Persistent memory for AI assistants via MCP

Adaptive Recall provides persistent memory for AI assistants using the Model Context Protocol (MCP). It solves the problem of AI assistants forgetting context between sessions, enabling long-term personalized interactions.

1 engagement·1 source·hackernews
Sun, Jul 12, 2026, 09:08 PM
Community

Developer reports AI coding agent with persistent memory across cold reboots

A developer on Reddit reports that their AI coding agent retained full context—including decisions, boundaries, and past mistakes—across a complete PC shutdown and fresh terminal session. The agent continued mid-thought without re-explanation or warm-up, suggesting a breakthrough in long-term memory persistence for coding assistants.

3 engagement·1 source·reddit
Sat, Jul 11, 2026, 01:56 PM
Incident

User reports Sonnet 5 bug causing prompt contamination and shares workaround

A user published an article detailing failure modes of Anthropic's Sonnet 5 model and an instruction layer to mitigate them. They also provide a temporary fix for a bug where user preferences/project instructions bleed into each prompt, contaminating the interaction loop. The user believes this bug is responsible for much of the initial negative feedback about Sonnet 5.

4 engagement·1 source·reddit
Sat, Jul 11, 2026, 05:20 PM