Import history
Every import is a tracked job with a preview, live progress, a completion summary, and recovery options. Nothing about an import is a black box.
Job lifecycle
flowchart LR
P[Pending] --> R[Running]
R --> D[Done]
R --> F[Failed]
R --> I[Interrupted]
F -- retry --> R
I -- resume or retry --> R
Pending — created after preview, waiting to run.
Running — articles convert in the background; progress is visible as total/completed/failed counts.
Done — a summary snapshot is stored with the job.
Failed — an infrastructure error stopped the job; retry continues from progress.
Interrupted — a server restart mid-import; the job can resume or be retried.
The preview step
Before any content is created, the preview returns what will happen: collections found, article counts per collection, target space options, and (for Nextra) the detected site root. Nothing is written during preview.
During the run
Live progress — see totals, completed, and failed counts while the job runs.
Failures are per-article — one bad article doesn't stop the import; failures are recorded (article, title, error) and listed at the end.
Cancel — running jobs can be cancelled.
When it finishes
The summary reports: collections created, articles published vs drafted, uncategorized articles, articles with conversion warnings, HTML block fallbacks, image rewrite failures, normalized note blocks, broken links, unsupported components, and asset rewrite failures.
Also available: the redirect map — source URL → new article URL for every imported article.
History and re-runs
Import history lists past jobs with their source, dates, and outcomes.
Retry a failed job (continues where it left off).
Reconvert re-runs the converter for a completed job — useful after converter improvements — without re-fetching the source.
Permission
Running imports requires a dedicated import permission (held by admins). Regular members can paste markdown or upload .md/.docx into the editor for one-off content without it.
Was this article helpful?