Short answer: yes. In Prefect 3.x, artifact keys are workspace-scoped, not per-flow-run. Any flow run can retrieve an artifact created by another flow run as long as they share the same workspace and you know the artifact’s key. What to know: - Keys are global within a workspace. They are not scoped to a flow, deployment, or run. - Creating a new artifact with an existing key creates a new version; it does not overwrite previous versions (except progress artifacts, which update in place). - Retrieving by key returns the latest version for that key. Example: create in one flow run, read in another

from prefect.artifacts import create_markdown_artifact, Artifact

# Flow A: create or update an artifact
create_markdown_artifact(
    markdown="# Daily summary\nAll systems go.",
    key="reports/daily-summary",
    description="Daily ops report"
)

# Flow B (or any other run): read the latest artifact by key
artifact = Artifact.get(key="reports/daily-summary")
if artifact:
    print("Artifact type:", artifact.type)
    print("Updated:", artifact.updated)
    print("Data:", artifact.data)  # for markdown, this is the markdown string
else:
    print("No artifact found for that key")

CLI to inspect versions

prefect artifact inspect "reports/daily-summary"

If you need a specific run’s artifact (not just the latest): - Easiest pattern: encode the run identifier (or a timestamp) into the key when writing, then read that exact key later. - Alternatively, you can query artifacts by

flow_run_id

via the REST API or the Python client filters. Docs and references: - Concepts: Artifacts concepts - How-to: Artifacts how-to - Python API: prefect.artifacts API - CLI reference: prefect artifact CLI