Data and Storage
Builders separates durable files, run-local files, configuration values, database connections, and private model connections. Choose the resource by its lifetime and ownership boundary rather than using one store for every kind of data.
Choose the right resource
| Resource | Lifetime | Best for | Accessed from |
|---|---|---|---|
| Static Drive | Persists across runs | Templates, exports, shared reference files, and durable workflow artifacts | Drive UI, Read File, Save File, and file-aware blocks |
| Context Drive | One workflow run | Uploads, generated intermediate files, and handoff between blocks | Read File, Save File, file-aware blocks, and the Code runtime |
| General Vault | Persists until changed or deleted | Reusable simple, JSON, dotenv-style, and secure configuration values | Template expressions and the workflow executor |
| Database connection | Persists until changed or deleted | Reusable PostgreSQL, MySQL, or MongoDB access | DB Query blocks in projects with matching ownership access |
| Private AI agent connection | Persists until changed or deleted | Provider credentials, model selection, and generation defaults | AI Agent blocks in the connection's ownership scope |
| External storage integration | Provider-controlled | Durable files that must remain in Google Drive, OneDrive, S3, or compatible storage | Read File and Save File through an integration |
Use Context Drive when a file only coordinates steps in the same run. Use Static Drive or an external provider when another run, user, or system must retrieve it later. Use General Vault for configuration values, not as a file system or database.
Keep data and credentials separate
A workflow may pass business data through block inputs and Context Drive. Credentials remain in a database connection, AI connection, General Vault secure entry, or provider integration. Do not paste secrets into labels, templates, Code blocks, workflow exports, logs, or screenshots.
Personal resources are available only where the current user and project scope allow them. Team resources support shared workflows, but access still depends on team membership and the selected project owner. Moving or importing a workflow never grants access to resources from its original scope.
Plan a data path
Before building the workflow, answer these questions:
- Must the data survive the current run?
- Is it a file, a configuration value, or a record managed by a database?
- Does one user own it, or must a team maintain it?
- Which block needs read or write access?
- Does an external provider require a dedicated integration?
- What should happen when the resource is missing, full, unavailable, or no longer shared?
Documentation map
- Static Drive covers persistent Builders files, folders, quota, previews, editing, and workflow paths.
- Context Drive covers temporary per-run files and the canonical File handoff.
- General Vault covers typed values, template references, secure entries, and rotation.
- Database Connections covers PostgreSQL, MySQL, MongoDB, TLS, ownership, and validation.
- Private AI Agent Connections covers providers, model discovery, defaults, capabilities, and provider billing.
- Resource Ownership and Workflow Visibility explains personal and team boundaries, transfers, imports, and repair.
