Artifacts & Storage

Artifacts & Storage is how you retrieve the files a job produces — PDB structures, SDF molecules, CSV reports — once it completes. Reach for it after a job you submitted through the Jobs API finishes and you need its outputs. It's the retrieval half of the job lifecycle that the Jobs API starts.

Artifacts & Storage

A finished job's outputs (PDB structures, SDF molecules, CSV reports) are stored with the job. GET /api/v1/jobs/:jobId (reference) returns the finished record, including its output artifact references.

Downloading Outputs

GET /api/v1/jobs/:jobId/artifacts (reference) lists a finished job's output files. Each entry has a download_url that works for 5 minutes. Fetch it with a plain GET and no auth header:

curl "https://www.ivybiosciences.com/api/v1/jobs/$JOB_ID/artifacts?types=pdb,csv" \
  -H "Authorization: Bearer ivy_your_key"

# then, for each artifact
curl -o structure.pdb "$DOWNLOAD_URL"

The route needs an API key or a signed-in session that can read the job. It returns 409 until the job has completed. A job you can't read returns 404, the same as a job that doesn't exist. When the links expire, call the route again for new ones.

The CLI does the same in one step:

ivybloom jobs download job_abc123

Files your chat's code sandbox exports have their own download operation: GET /api/harness/sandbox/:chatId/export/:assetId/download (reference) returns a short-lived signed link.

Artifact Types

The download route lists these types; pass any subset as types:

Without a key

The operations on this page need an API key or a session. A caller with no key starts with the keyless orchestrate request in Guest curl; guest behaviour and limits are stated once, in llms.txt.

Read next

• Jobs API — job submission and status
• Guest curl — the keyless first call
• SDK Artifacts — TypeScript artifact helper