Developer onboarding

From API key to first artifact.

A practical first-run path for developers building around DocuZeus jobs, public samples, and repeatable processing.

First run shell

export DOCUZEUS_API="http://localhost:8000"
export DOCUZEUS_API_KEY="docz_dev_key"

curl -X POST "$DOCUZEUS_API/v1/documents/upload" \
  -H "Authorization: Bearer $DOCUZEUS_API_KEY" \
  -F "[email protected]"

curl -X POST "$DOCUZEUS_API/v1/processing/compress" \
  -H "Authorization: Bearer $DOCUZEUS_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: first-run-DOC_ID" \
  -d '{"document_id":"DOC_ID","quality":"medium"}'

curl "$DOCUZEUS_API/v1/jobs/JOB_ID/status?wait=10" \
  -H "Authorization: Bearer $DOCUZEUS_API_KEY"

Explore the workflow

Developer Onboarding

Create a key and run the first artifact.

Happy path

Start, build, verify, evaluate

Pipeline tool Explore the workflow

Start

Create a scoped key, upload a public sample, submit one job, and keep the returned IDs.

Open onboarding 2

Build

Choose a workflow, inspect route families, then copy the TypeScript, Python, or curl recipe.

Open recipes 3

Verify

Check job status, generated artifacts, and proof artifacts before handoff.

Open proof 4

Evaluate

Review examples, capability evidence, and security guidance before release.

Open guide

Integration readiness

Keys

Scoped key and CSRF path confirmed.

Jobs

Submit, poll, retry, and timeout states wired.

Artifacts

Download URLs and manifests checked.

Alerts

Queued, failed, and complete feedback visible.

First-run checklist

Use this tighter loop for the first command-line pass, then graduate the same route sequence into the broader workflow path above.

Create a scoped key

Start with documents:read/write, processing:run, and jobs:read for a basic processing integration.

Open step 2

Upload a public sample

Use Form W-9 or the NIST framework example so the test path does not depend on private data.

Open step 3

Submit a job

Run OCR, compress, preflight, compare, or a saved pipeline through the canonical /v1 route.

Open step 4

Poll with backoff

Use job status endpoints, terminal statuses, and idempotency keys so retries do not duplicate work.

Open step 5

Download or import artifacts

Fetch the result PDF, CSV, JSON report, or evidence bundle and keep it tied to the source workflow.

Open step

Next run

After the first artifact works, move the same sequence into a saved or ad hoc pipeline, then use language recipes for retry, polling, and artifact download code.

Integration overview Integration overview Integration overview