AIStable3 steps

Prepare Data for AI

Normalize records and strip personal identifiers (emails, phones) so a dataset is safe and consistent to feed into an AI model or embedding pipeline.

How it works

Pipeline

Normalize for consistency, then redact contact PII before the data leaves your system.

  1. Transform
    • Trim stray whitespace from every field.
  2. Normalize
    • Standardize "email" into a consistent email format.
    • Standardize "phone" into a consistent phone format.
  3. Redact
    • Replace detected email and phone with readable labels like [EMAIL].

Frequently asked

FAQ
Is this recipe ready to run?
Yes. This recipe is stable — every step maps to a shipped capability, so it runs and returns structured records today.
Can I customize the steps?
Yes. A recipe is a starting configuration. Install it into a project and you can add, remove, or reorder steps to match your own data.
What does it cost to run?
Recipes run on the same usage-based pricing as every capability. Start free with 1,000 usage units a month across everything — no credit card required.
How do I run it?
Install the recipe into a project from your dashboard to get a saved pipeline, then call it over the REST API, an SDK, or MCP — one request in, clean records out.
prepare-data-for-ai · 3 steps · Stable

Run this recipe on your own data

Start free with 1,000 usage units a month across every capability — no credit card. Install the recipe, adjust the steps, and call it in one request.

Make data consistent and safe before it reaches a model — and tag it for triage.