Glideworth never connects to your bank, so your numbers arrive by file or from a script you control. Every importer is built around the same rule: show what would change before changing anything.

Importing

ProjectionLab JSON. Export your data from ProjectionLab (account menu, JSON export), then drop the file into the app’s Import screen or send it to POST /api/v1/import with "format": "projectionlab". The import report lists every key by path, with per-entity counts, and sorts them into what mapped exactly, what was approximated and what has no equivalent. It never echoes your values. Anything marked approximated or unsupported needs a manual check. A ProjectionLab plugin API key found in the file is discarded. Re-importing an existing household needs an explicit mode: create (the default), replace, or merge.

CSV. Three kinds: account balances, progress history, and transactions (summarised into annual expense events on a plan). Presets cover Monarch and YNAB exports, and a generic mode detects the delimiter, columns and number formats. For balances you see a reconciliation (matched, created and unmatched accounts with before-and-after balances) before applying. The dry run is the default; applying is atomic, so finances, snapshots and the plan change in one transaction or not at all.

OFX, QFX and QIF. Bank, card and investment statements. Balances reconcile like a CSV balance import. Spending is summarised into annual expense events, with transfers, income and card payments left out. Account numbers are masked to their last four digits. OFX 1.x and 2.x are both read.

Glideworth’s own export, to move a household between instances, between a self-hosted server and (once it opens) the hosted service, or back from a backup.

A dry-run report, as a person would read it (counts are illustrative):

Result Entities Example
Mapped exactly 41 Accounts, balances, salary
Approximated 6 A growth schedule with no equal in Glideworth
No equivalent 2 A chart preference

Worked illustration

You export a household from another planner and a month of statements from your bank. Dry-run the planner file: 41 items map exactly, 6 are approximated and 2 have no equivalent, each listed by key path. You fix the approximated growth schedule in the plan editor, then apply the import. Next, preview the bank file: the report says three accounts matched, one debt was not found, and the savings balance differs by £212, with the before and after shown. You map the debt, apply, and a progress snapshot is recorded. Counts are illustrative.

Exporting

  • The household document as JSON: household, current finances, plans, progress and settings, in a schema-versioned file that Glideworth can import again. Older files are migrated when read.
  • Tables as CSV or JSON: year by year, cash flow, taxes and account balances, in the basis you choose, plus progress history.
  • A report with key figures, milestones, assumptions, current finances and standing caveats, as data and as a printable view. Your browser’s print function makes the PDF.
  • A whole-database backup is separate: see self-hosting.

The importers and exporters are a separate package under the MIT licence, so other software can read and write these formats. See open source.

What it does not do

  • No live sync with budgeting apps. Balance sync with Lunch Money, YNAB, Monarch and Actual is on the roadmap; today it is a script, CSV or a bank file.
  • No bank aggregation, by design.
  • ProjectionLab is the only other planner format read directly. There are no importers for other planners.
  • Imports are approximate where formats differ. The report says where; check those items.
  • Table export covers four tables, not any chart you can draw.
  • No server-side PDF. Print the report from the browser.
  • The JSON export is not a full backup: it leaves out users, API keys and version history.

Go further

See where your plan lands.

Free to self-host today, with every feature. The hosted service opens later, at $9 a month or $90 a year.