Cmd. /memories
Configure memory use and generation for Codex sessions.
Key takeaways
/memoriescontrols whether Codex uses existing memories and whether new sessions contribute to memory generation.- Disable memory behavior before working with sensitive external context, such as client-specific work.
- Unlike
/compact, which only summarizes the current transcript,/memoriesgoverns long-term context boundaries. - Be conservative when web, MCP, or browser context is involved, and treat memories as a convenience layer with explicit privacy rules.
Original Command Declaration
/memoriesOfficial Summary
/memories controls whether Codex uses existing memories and whether new sessions contribute to memory generation.
Usage
/memoriesGood Examples
- Disable memory behavior before working with sensitive external context.
- Verify memory settings when a long-running project depends on team preferences.
Similar Commands
| Command | Difference | Choose It When |
|---|---|---|
/memories | Controls memory use and generation | Long-term context boundaries matter |
/compact | Summarizes the current transcript | You only need to reduce current context |
/clear | Starts fresh visible chat flow | You want a new conversation surface |
Use Cases
- Keeping client-specific work out of memory generation.
- Auditing whether memory settings match team policy.
Cautions
- Treat memories as a convenience layer with explicit privacy rules.
- Be conservative when web, MCP, or browser context is involved.
Release Basis
- Judgment: official docs listing
- First evidence version: official slash commands documentation
- Date (UTC): verified 2026-05-21
- Release link: https://developers.openai.com/codex/cli/slash-commands
Sources
- Codex slash commands: https://developers.openai.com/codex/cli/slash-commands
- Codex memories: https://developers.openai.com/codex/memories