How do I verify Montage workspace access with JetBrains AI Assistant?
JetBrains AI Assistant has its own Model Context Protocol settings, separate from the Copilot extension and from Junie. It can register remote servers with a JSON URL entry and choose a global or project server level. Use that separation to keep Montage access personal when appropriate, then test execution explicitly instead of assuming that applying IDE settings completes remote workspace authorization.
Using JetBrains AI Assistant, call only Montage signals_status as the first live operation. Report the actual workspace identity and coverage from the tool response. If authentication, transport, or the tool is unavailable, stop and explain the blocker. Do not start collection, recurring tracking, writes, publishing, messaging, or social actions.Open AI Assistant’s MCP settings
Go to Settings, Tools, AI Assistant, Model Context Protocol (MCP), then choose Add. Inspect the table first for an existing Montage entry. The documented Import from Claude action imports configurations, not proof of consent transfer. Select the current remote registration path rather than a local working-directory example, since a hosted service has no executable for the IDE to launch.
Provide the remote JSON shape
Use an mcpServers object containing a montage entry with url set to https://app.montages.co/mcp. Choose the server level deliberately: global availability follows you across projects, while project scope limits where the definition appears. Keep credentials out of the JSON sample. A saved URL identifies a network destination but neither signs in to Montage nor authorizes the model to use every available tool.
Apply without enabling everything automatically
Review the Automatically enable new and changed MCP servers setting before clicking Apply. It controls startup behavior, not Montage permissions. Monitor the Status column to distinguish a server that never started from one that initialized successfully. Avoid changing this preference globally just to make one entry appear active; other configured services may be affected by the same IDE-wide behavior.
Check the installed authentication behavior
The remote URL examples establish connection configuration, not a universal Montage OAuth guarantee across IDE releases. Inspect your version's supported authentication path and complete consent only when compatible. Otherwise use an existing owner-approved credential through supported secure settings or choose another client. The account paying for AI Assistant does not supply a Montage token or an automatic grant to a Montage workspace.
Use a named read-only tool
Inspect the tools through the Status column control, then explicitly invoke signals_status in AI Assistant as the first live Montage operation. Review the actual result for the intended workspace. For failures, use Reconnect or the documented mcp logs directory according to the error, without logging credentials. Do not turn a successful IDE connection into unattended research, collection, or social actions.
Keep the boundaries clear.
- IDE server levels and remote authentication options depend on the installed AI Assistant release.
- Documented setup, not live interoperability certification. A visible tool catalog is not workspace authorization.
- First live operation: read-only signals_status. No automatic collection, recurring tracking, or social actions; broader actions need separate explicit authorization.
Start with a read-only check: ask your agent to report actual source coverage before collecting records or exporting an audience.
Primary references
Documentation reviewed 2026-11-15. Product capabilities, client availability and plans can change. Third-party marks identify their products, not an affiliation.