GitHub Copilot for JetBrains Adds Memory and Ollama Support

GitHub shipped a substantial update to GitHub Copilot for JetBrains, headlined by a new Copilot Memory feature and bring-your-own-key support for Ollama as a model provider, alongside expanded enterprise controls.

Details

• Copilot Memory: The plugin can now retain and recall useful information across agent chat sessions, so users don’t need to repeat project context in every new conversation; the feature can be toggled from the Copilot settings portal
• Ollama as a BYOK provider: Developers can now configure Ollama as a bring-your-own-key model provider, with provider configuration and model selection available throughout the JetBrains experience
• Enterprise server-based controls: Administrators get new server-based controls for managing Copilot across their organization, covering plugin availability, MCP server access, and related settings
• Automatic CLI installation: The plugin can now automatically install the Copilot CLI from integrated terminals on macOS, Linux, and Windows
• Codex workflow visibility: Codex sessions are now visible in debug logs and support updated permissions and customization options
• Usability fixes: The update also improves account switching/removal, the model picker and settings panel, restores file/folder references in chat inputs, repositions customization and debug log buttons, and fixes issues with MCP execution reliability, terminal output handling, cloud agent stability, and diff-based editing

How to try it

The update is available now to all GitHub Copilot for JetBrains users; Copilot Memory can be enabled or disabled from the Copilot settings portal, and Ollama can be added as a model provider from the same provider configuration screen.