Sharing Data Between Jobs With Artifacts

Jobs run on separate machines, so to pass a build from one job to another you upload it as an artifact.

Because each job gets a fresh, isolated runner, a file created in one job does not exist in the next. Artifacts solve this: you upload files in one job and download them in another. This lesson shows the upload/download pattern and when to use it.

The problem artifacts solve

Imagine a build job that compiles your app and a separate deploy job that ships it. The deploy job starts on a clean runner with none of the build output. Artifacts bridge that gap: the build job uploads its output, and the deploy job downloads it, even though they ran on different machines.

Uploading an artifact

Use actions/upload-artifact to store files under a name. Anything in the given path is bundled and kept by GitHub for later retrieval.

.github/workflows/build.yml

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - run: npm ci && npm run build
      - uses: actions/upload-artifact@v4
        with:
          name: dist
          path: dist/

Downloading it in a later job

A downstream job uses needs to wait for the build, then actions/download-artifact to pull the files in. The downloaded files appear on its runner exactly as uploaded.

downstream job

  deploy:
    needs: build
    runs-on: ubuntu-latest
    steps:
      - uses: actions/download-artifact@v4
        with:
          name: dist
          path: dist/
      - run: ./deploy.sh dist/

Things to keep in mind

Artifacts are also great for keeping test reports, logs, and coverage files for inspection.
They count against storage and have a retention period - set retention-days to expire them sooner.
For passing a single small value (not a file) between jobs, use job outputs instead of an artifact.
Large artifacts take time to upload and download; only ship what the next job actually needs.

Key takeaways

Artifacts move files between isolated jobs via upload and download.
Pair upload-artifact/download-artifact with needs to order the jobs.
Use job outputs for small values, and set retention to control storage.