Redact personal data from every transcript and ticket

Where personal data redaction usually starts

Most teams redact with patterns for emails and phone numbers, then add a frontier model for names and addresses that patterns miss. Patterns break on new formats, and the model call costs the same on every line of every transcript. A Facet model applies your rules to every text in one call.

How it works

1. Describe the task. Paste the prompt you use today, or write it plainly: “Replace names, emails, phone numbers and street addresses with [NAME], [EMAIL], [PHONE] and [ADDRESS]. Keep company and product names.” Add a few real examples if you have them.
2. Agree on what good looks like. Facet writes plain statements of a good answer, like “an email address never survives, even when it is written out as dana at example dot com”, each with examples. Edit any that are off.
3. Create, then correct. Facet builds the model on its own, in the background, and tests it against every statement before it goes live. When a detail slips through, say what should change. Facet quotes the fix before it spends anything.

Two ways to use it

• From your agent: connect Claude, Cursor, ChatGPT, Codex or VS Code to Facet, then paste:

Use Facet to build a model that replaces names, emails, phone numbers and street
addresses in support transcripts with placeholders like [NAME]. Here are ten real
transcripts: ... Show me the draft and the price before you approve anything.

• In your code: keep your OpenAI SDK and change two lines, the address and the model name.

import os
from openai import OpenAI

client = OpenAI(base_url=os.environ["FACET_BASE_URL"], api_key=os.environ["FACET_API_KEY"])
reply = client.chat.completions.create(model="redact-pii", messages=[{"role": "user", "content": transcript}])

Why Facet for personal data redaction

• Every rule, every line. Your masking rules hold the same way on the millionth line as on the first.
• Text stays out of logs. Facet stores call details such as token counts, never the text your app sends.
• Checked before every update. An update goes live only if it keeps every statement the live version passes, and any update rolls back in one step.
• Yours. Export it any time, free, and run it wherever you like.