[Idea] Import from Cospend

[eikaramba]

Affected area

Import and export

Problem or opportunity

I'm considering switching from cospend to spliit-cloud, however i did not see an import from cospend. given the popularity of cospend i think it makes sense to support it.
https://github.com/julien-nc/cospend-nc

Desired outcome

import projects & bills from cospend along with recurring bills and match users

Alternatives or workarounds

No response

Additional context

No response

Before submitting

• I searched existing issues for similar ideas.
• I removed personal, financial, and secret information from this public proposal.

[pullfrog]

Summary

The author wants to switch from Cospend (the Nextcloud shared-expense app) to Spliit Cloud and asks for a one-time importer: Cospend projects → Spliit groups, bills → expenses, recurring bills → recurring series, and Cospend members matched to Spliit users. The underlying goal is the standard switcher problem — without an import path, migrating means losing all history. This fits the existing "Migrations & imports" roadmap direction (Splitwise CSV shipped, Tricount planned), and no duplicate issue exists.

Feasibility

High. The import system is already built for multiple providers — Spliit JSON/CSV/URL, Spliit Cloud ZIP, and Splitwise CSV all feed the same pipeline, with Tricount/Settle Up placeholders as stubs for exactly this pattern. Participant mapping (name-based auto-match, invite by link/email, unlinked), category mapping, currency conversion, and collapsing recurring schedules into Spliit's recurring-expense series are all provider-agnostic and already exercised. A Cospend importer is essentially a new parser for Cospend's export plus a wizard tab.

Limitations:

• Cospend members are names + weights only (no emails), so "match users" can only be name-based auto-match plus manual assignment or invite-by-link — email matching isn't possible from Cospend data alone.
• Some Cospend metadata (payment modes, custom categories, bill attachments) has no Spliit equivalent and would be dropped or degraded.
• A sample Cospend export from the author (JSON vs CSV) would de-risk the parser work.

Scope

Medium. Mostly additive — a new source parser and wizard wiring that reuses the existing mapping, matching, recurrence, and creation machinery. It spans the parser, import UI, and i18n, but touches no core data model, permissions, or existing import behavior.

  ^{｜ View workflow run ｜ via Pullfrog ｜ Using openai-compatible/omen-alpha ｜ 𝕏}

[antonio-ivanovski]

I am all for this! I see that you are familiar with coding, so would really appreciate you opening PR with working Cospend importer, since there are already couple of importers, any coding AI shall be capable of implementing this.

[eikaramba]

done, i have finished the import from cospend. i used real data from my cospend instance (with complex bills, multiple users and different recurring bills) and also have checked everything manually again after the automatic tests passed as well. looks quite good and seems to work. i also checked the code briefly myself and it looks good. explanations from ai are also good. LGTM :)