Your archive, explained.

Changing your email and recovering access

Open Account → Email address to change your sign-in email. You'll need a recent sign-in and a code from each mailbox. After confirmation, every session signs out and your old address receives a notice. Your backups, connections and subscription stay with the same account.

If you use Google sign-in, the linked Google account continues to work. Changing your KeepMyBase email does not change that Google account or your Google Drive connection.

If you have lost mailbox access, try your already-linked Google account or your email provider's recovery process. Changing the email still requires both mailboxes. We cannot transfer account ownership based on billing details or access to an archive. Contact help@keepmybase.com for help identifying your recovery options.

Your saved archives remain accessible in your own storage even if you cannot sign in to KeepMyBase.

What gets captured

The archive format includes table records in CSV and JSON, table and field metadata, original record IDs, actual supported attachment files, and a manifest with checksums.

Airtable is read over a capture interval. Changes made during a run can appear at different points in different tables. The manifest records the capture interval.

Interfaces, automations, extensions, forms, sharing permissions, comments, revision history, and exact view configuration remain in Airtable. Recovery uses the exported records, attachment files, and their source mappings.

Reading your files

records.csv

Open in Excel, Google Sheets, or another spreadsheet. Values that could be read as formulas are escaped in the export format.

records.json

Structured records with their original Airtable IDs. Useful for scripts and re-import preparation.

attachments/

The downloaded files, organized by attachment ID. The manifest links them to their original records.

schema.json

The captured table names, field names, field types, and IDs.

manifest.json

Capture timestamps, record counts, and checksummed table indexes. Each index lists the export parts and their file checksums.

Tables are split into numbered parts. Keep every part when moving or recovering an archive. Each CSV has its own header; JSON preserves the source IDs needed to reconnect linked records. Earlier archives use a single export per table.

The sample contains 3 fictional records and one text attachment.

Recovering a base

Start with a complete archive from your storage and an empty Airtable base. Keep the original archive while you check the recovered data.

1. Download the entire archive folder, including every export part and attachment. Use the preparation tool below to check file lengths, checksums and record links.
2. Import the prepared CSV parts into their matching tables. Turn off automatic field type selection and keep the source record ID as each table’s primary field. Append later parts to the same table once. For each appended part, check Map Fields and enable Other settings → Exclude first row in import. Header-only CSVs preserve empty tables.
3. After importing every table, convert linked-record columns using the table mappings in recovery.json. Select the existing imported table; Airtable matches the saved source IDs. Rebuild one direction of each relationship and check its automatically created reciprocal field.
4. Restore field types and rebuild computed fields using the saved schema. Switch primary fields back after links resolve. Keep source IDs for verification.
5. Upload the saved attachments using the prepared attachment maps. Check each file’s record, field and original filename, including files with identical names.
6. Compare record counts, links, empty values, checkboxes, dates, computed results and attachment bytes before using the recovered base. Airtable assigns new record IDs and system timestamps. Recreate views, automations and permissions separately.

Saved formulas can reference original field IDs. Replace those references with the corresponding field names. Airtable’s API omits some computed-field settings, including rollup aggregation formulas; reconstruct those rules and compare their results with the captured values. Collaborator fields need manual identity mapping.

Prepare the import files locally

The tool runs with Node.js 24 on your computer. It reads a downloaded archive and creates a separate recovery folder. It requires no account, credentials or network connection. Your original archive stays unchanged.

Run node prepare-recovery.mjs "/path/to/archive" "/path/to/new-recovery". The new folder includes import CSVs, attachment maps and step-by-step instructions.

Prepared CSVs preserve original text, including values that spreadsheet software may evaluate as formulas. Import the records directly into Airtable and read attachment maps in a text editor. The archive’s original records.csv files remain the safer choice for spreadsheet viewing.

When a capture needs attention

A complete capture has saved every item within the declared coverage. An incomplete capture means some items were missed. A failed capture did not finish the archive.

Review the run details first. An expired attachment link needs a fresh capture. An expired connection needs reconnection. A full destination needs available storage. A temporary provider outage can be retried.

Your last complete archive stays visible alongside the latest run's result.

Temporary files from failed or cancelled attempts are eligible for cleanup after 24 hours. Cleanup requires storage access and confirmation that the capture has stopped. Completed archives follow your retention setting.

Retention and ownership

Choose how long to keep older captures, up to the plan's retention limit. Your latest complete archive remains available.

Your archives live in your own storage. Removing a schedule or canceling your subscription preserves your completed files in that storage.

Connections and sign-in

KeepMyBase uses email sign-in codes and optional Google sign-in. A single account can organize multiple Airtable and storage connections. Each backup has its own source and destination.

Manage Airtable, Google Drive and S3-compatible storage from Connections. Run your first backup, then enable daily backups and retention in its settings.

Connect an S3 bucket

Create a private bucket in your storage provider, then choose Add storage destination in Connections. Select the provider and copy its S3 endpoint, bucket name, and region. Give the connection a recognizable name and use a dedicated folder prefix such as keepmybase, without a trailing slash.

Amazon S3

Use the bucket's region and its matching regional S3 endpoint.

Cloudflare R2

Copy the account's S3 API endpoint, including jurisdiction. The region is auto.

Wasabi, B2, Spaces

Copy the provider's regional S3 endpoint and use that same signing region.

Local MinIO testing

The local development fixture uses localhost:59000. Hosted custom endpoints require additional network isolation before release.

Create a key restricted to the dedicated bucket or prefix with list, read, write, and delete permission. Enter its access key ID and secret access key in the form. KeepMyBase encrypts the saved keys and never displays them again. Path-style addressing is selected by default; leave it selected unless your provider requires virtual-host addressing.

Test and save connection writes a unique small file inside the prefix, reads and checks it, lists it, then removes it. If the test fails, check the endpoint, region, bucket, credentials, and permissions. A failed cleanup can leave a small .connection-test file; versioned buckets can retain old versions of test files.

To rotate a key, select Replace credentials. A successful test replaces the saved key; a failed test keeps the previous credentials. You may be asked to confirm your account with a fresh email code. Disconnect removes the saved connection and its keys while leaving your bucket files in place.

Provider instructions: AWS permissions, R2 bucket tokens. For help, email help@keepmybase.com.