notebooklm-mcp-cli-auth-format-two-files
Upstream notebooklm-mcp-cli uses TWO auth files, not one: profiles/default/cookies.json (cookies) and profiles/default/metadata.json (session metadata). Our custom wrapper used a single storage_state.json in notebooklm-py format. CSRF token and session_id are auto-extracted by upstream on first API call — only cookies are required for migration.
Related
- notebooklm-mcp-cli-auth-two-file-format
- docker
- notebooklm-mcp-cli-consolidated-tool-pattern
- notebooklm
- notebooklm-custom-wrapper-six-missing-capabilities
- notebooklm-upstream-notebook-get-replaces-source-list
- notebooklm-upstream-container-memory-1g-minimum
- notebooklm-mcp-cli-env-var-config-path
- notebooklm-cookie-refresh-cron-must-update-on-migration
- notebooklm-py-cookies-rejected-by-upstream-client-fingerprin
- notebooklm-mcp-cli-auth-two-file-profile-format
- notebooklm-mcp-cli-auth-requires-fresh-capture-not-migration
- notebooklm-upstream-auth-fingerprint-rejection
- notebooklm-upstream-auth-google-fingerprint-rejection
- notebooklm-upstream-auth-two-file-format
- chrome-devtools-console-paste-protection-requires-allow-past
- google-auth-cookies-httponly-use-network-tab-for-extraction
- notebooklm-mcp-cli-auth-requires-two-files-not-one
- browser-httponly-cookies-not-accessible-via-document-cookie
- notebooklm-upstream-auth-format-two-files
- upstream-notebooklm-auth-needs-two-files-not-one
- notebooklm-upstream-auth-uses-two-file-profile-format
- mc-dispatcher-auth-header-x-api-key-not-bearer