packages/twenty-docs/user-guide/data-migration/how-tos/migrating-from-self-hosted-to-cloud.mdx
This guide walks you through migrating your data from a Twenty self-hosted instance to Twenty Cloud. The process involves setting up your cloud workspace, exporting your data, and re-importing it.
<Note>Views, workflows, and roles must be recreated manually after migration. Plan time for this configuration work.</Note>
Before importing data, recreate your custom objects and fields:
<Note>Take screenshots of your self-hosted data model for reference, or keep both instances open side by side.</Note>
Users must accept their invitations before you import any records that reference them (like Account Owner fields). If users don't exist yet, those relations cannot be mapped. </Warning>
Export each object from your self-hosted instance:
companies-export.csv)Export in this order (for reference when importing):
The exported CSVs contain user IDs from your self-hosted instance. These IDs won't match your cloud instance, so you need to replace them with emails.
For each CSV file with user references (Owner, Assignee, etc.):
accountOwnerEmail next to accountOwnerId)Example:
Before:
name,domain,accountOwnerId
Acme Corp,https://acme.com,old-uuid-123
After:
name,domain,accountOwnerEmail
Acme Corp,https://acme.com,[email protected]
<Note>Use the same email addresses that users used to accept their cloud workspace invitation.</Note>
Import files in the correct order to maintain relationships:
See How to Import Relations for details on maintaining relationships.
For each CSV file, in order:
After importing data, manually recreate:
<Check>All data imported successfully</Check> <Check>Relations between objects working correctly</Check> <Check>User assignments (Owner, Assignee) mapped correctly</Check> <Check>Views recreated</Check> <Check>Workflows recreated and tested</Check> <Check>Roles and permissions configured</Check> <Check>Email/calendar sync reconnected</Check> <Check>API integrations updated with new keys</Check>
For complex migrations or large datasets, contact us at [email protected] or explore our Implementation Services.