Migrating from Harvest

If you are coming from Harvest, Time Trakkr can bring your data across so you keep your history and start where you left off. The migration is an admin task and lives on the Integrations page, in the Migrate from Harvest card. It runs in two steps, and both are safe to run more than once.

Connecting to Harvest

The card's badge reads Not connected until you link your account.

1. Open Integrations and find Migrate from Harvest.
2. Enter your Harvest Account ID.
3. Enter a Harvest Personal access token.
4. Select Connect. While it checks, the button reads "Checking…" and, once linked, the badge shows Connected with your Harvest company name.

You can select Disconnect at any time to unlink.

Step 1: import clients, projects, tasks, and people

Select 1 · Import clients, projects, tasks & people. While it runs the button reads "Importing…".

This brings over your clients, projects, tasks, and people. How people arrive depends on their Harvest status:

• Active Harvest people are created as invited members. Their record, including rates and capacity, is imported, but they cannot sign in until you send them an invitation from Team and they accept. The results list them so you know who to invite.
• Archived (inactive) Harvest people are created as disabled members. They cannot sign in and are left out of active workflows, but their past time entries still attach, so your historical reporting and profitability stay complete.

Time Trakkr matches Harvest people to existing members by email, so anyone already in your workspace is updated rather than duplicated.

Step 2: backfill time entries

Select 2 · Backfill time entries to bring in your full time history. While it runs the button reads "Backfilling…". This step runs in the background and can take a while for large accounts; the card shows the status as it progresses.

There is an optional only before date. It defaults to your first day tracked in Time Trakkr, so imported Harvest history and hours you have already logged here never double-count. Clear the date to import everything.

Re-running is safe

Both steps are built to be run again without creating duplicates. Clients, projects, and tasks are matched to their Harvest records and updated in place, and time entries that already exist are skipped. If an import is interrupted, or you add more in Harvest later, just run the step again.

What re-imports keep in sync

On a repeat run, Step 1 refreshes a few things from Harvest:

• Roles. A plain member is elevated to their Harvest access level (admin or manager). It never demotes anyone and never changes existing owner, admin, or manager assignments.
• Contractor status. Harvest's employee-versus-contractor classification is treated as authoritative, so this label is kept current even for people imported before the field existed. See roles and permissions for what that label means.
• Project auto-assign. Whether a person is automatically added to future projects is synced from Harvest.

Manual choices you have made in Time Trakkr, such as filled-in reporting lines and work roles, are left alone. Once your people are imported, invite the active ones from Team when you are ready for them to sign in.

What happens to edits you have made here

Re-running an import does not overwrite changes you made in Time Trakkr. The import compares three things: what it wrote last time, what the record says here now, and what Harvest says now. A field only changes if you have not touched it since the last import.

Where both have changed — you renamed a client here, and it was renamed in Harvest too — your version wins, and the disagreement is recorded rather than discarded.

Integrations → Kept your version lists those: which record, which field, your value and Harvest's, and when it was last seen. Nothing there is waiting on you; the decision has already been made the right way round. It is there so you can disagree:

• Keep mine records that you looked and agreed.
• Use the source value takes Harvest's value and stops that field conflicting in future, so it tracks Harvest again from then on.

A field that keeps disagreeing stays a single entry with the date it was last seen, so a weekly import does not pile up copies. If the same row keeps coming back, the two systems disagree about something worth settling in one of them.

Note this starts from the next import after the feature arrived. Records imported before then carry no snapshot of what was written, so they are treated as untouched — which is the previous behavior, and means the change cannot retroactively freeze a field nobody edited.