CasePatch API quickstart

Production data anonymization for AI agent evals, from Python or any HTTP client.

1. Install

pip install casepatch              # cloud client, no dependencies
pip install "casepatch[local]"     # adds the local engine (runs on your machine, nothing uploaded)
casepatch setup-local              # one-time model download (~210 MB) for local mode

2. Anonymize

Local (no key, nothing leaves your machine):

from casepatch import CasePatch
cp = CasePatch(mode="local")
cp.anonymize("statement.pdf", out="statement.anon.pdf")

Cloud: create a key on your account page (shown once), then:

export CASEPATCH_API_KEY=cpk_...
casepatch anonymize ./docs --out ./anon --cloud

3. REST

curl -X POST https://casepatch.com/api/v1/anonymize -H "Authorization: Bearer $CASEPATCH_API_KEY" \
     -H "Content-Type: application/pdf" --data-binary @statement.pdf            # -> {"id": "...", "status": "queued"}
curl https://casepatch.com/api/v1/jobs/$ID -H "Authorization: Bearer $CASEPATCH_API_KEY"   # status, summary, changes
curl -o anon.pdf https://casepatch.com/api/v1/jobs/$ID/result -H "Authorization: Bearer $CASEPATCH_API_KEY"
curl -X DELETE https://casepatch.com/api/v1/jobs/$ID -H "Authorization: Bearer $CASEPATCH_API_KEY"

OpenAPI: https://casepatch.com/api/v1/openapi.json.

Limits and data handling

Markdown version · OpenAPI spec