PushizoDocs
Analytics & data

Exports and imports (CSV)

CSV export jobs (users, subscriptions, events, deliveries), signed download links, Pushizo and legacy provider CSV imports, error reports and retention periods.

Bulk data transfer uses an asynchronous job model: you create a job, a worker produces/processes the file, and you poll its status as queued → running → completed | failed. Files are stored in S3-compatible storage under the org/{org_id}/app/{app_id}/… prefix; download links are signed URLs valid for 24 hours.

Exports

POST /v1/apps/{app_id}/exports { "type": "users" | "subscriptions" | "events" | "deliveries" | "message_events", "filters"?: { "from"?, "to"?, … } } (admin) → { id, status }. Status: GET …/exports/{id} → status, row_count, download_url and expires_at (once completed), error; list: GET …/exports?cursor=.

TypeContents
usersUser profile, aliases, tags/properties, language, time zone, created/last active
subscriptionsSubscriptions: channel, status, device, opt-in/opt-out times (tokens masked)
eventsEvents: name, properties, value, time
deliveriesid, message_id, campaign_id, subscription_id, user_id, channel, status, error_code, suppression_reason, sent_at, delivered_at, opened_at, clicked_at, created_at
message_eventsDelivery event timeline (sent/delivered/opened/clicked/bounce…)

Per-message CSV: GET /v1/notifications/{id}/history (deliveries) and POST …/export_events (events). Organization-level data export requests and audit log exports (a plan feature) are in the dashboard under Organization → Data & privacy; a daily NDJSON audit stream is available for SIEM (security).

Imports

POST /v1/apps/{app_id}/imports (multipart/form-data: type, file; admin):

typeFormat
csv_usersPushizo CSV: columns external_id, email?, phone?, push_token?, channel?, language?, timezone?, tags.<name>?, properties.<name>?; rows with the same external_id are merged
legacy_playersLegacy provider "Export players" CSV (migrating from a legacy provider)

Rules: UTF-8, first row is the header, at most 50 MB; emails are lowercased and phone numbers normalized to E.164, and invalid rows are written to the error report (GET …/imports/{id} → status, stats (processed/failed rows), error_report_url, file_name). Imports do not produce webhooks and are reflected in quota counters as subscribers. The uploaded file is deleted after IMPORT_FILE_RETENTION_DAYS (default 7 days).

cURL:

curl -X POST "https://api.pushizo.com/v1/apps/app_…/imports" \
  -H "Authorization: Bearer $PUSHIZO_API_KEY" \
  -F "type=csv_users" -F "file=@users.csv"

In the server SDKs, use hs.imports.create({ type, file }); examples in API → Exports / imports. Dashboard: Audience → Export/Import.

Operations

Notes on worker concurrency, memory and storage quota for large imports/exports: import/export runbook. Performance reference: exporting 100,000 users ≈ 2 s (18 MB CSV).

On this page