expunct-cli 0.2.0

CLI for the Expunct PII redaction API — redact, detect, and manage sensitive data from the command line.

Expunct CLI

Privacy infrastructure for modern applications.

Redact PII, secrets, and sensitive data from text, logs, and files — before it reaches AI, analytics, or external APIs.

---

🚀 Quick Start

pip install expunct-cli

export EXPUNCT_API_KEY=your_api_key

expunct redact --text "John Smith email [e-mail]"

Output:

PERSON_1 email EMAIL_1

---

✨ Why Expunct?

Modern applications constantly handle sensitive data:

• AI prompts sent to LLMs
• application logs
• customer support tickets
• analytics pipelines

Expunct helps you:

• 🔒 Detect PII (emails, phone numbers, names, etc.)
• 🧠 Detect secrets (API keys, tokens, credentials)
• 🧩 Apply policies (redact, mask, pseudonymize)
• ⚡ Sanitize data before it leaves your system

---

🧪 Examples

Redact sensitive data

expunct redact --text "Contact me at [e-mail]"

Detect entities without modifying text

expunct detect --text "My email is [e-mail]"

Process a file

expunct redact logs.txt

Redact binary files (PDF, DOCX, images, video, audio)

expunct redact report.pdf --output redacted_report.pdf

Use with pipes (great for scripts & agents)

cat logs.txt | expunct redact
echo "My SSN is 123-45-6789" | expunct detect

Cloud URI redaction

expunct redact --uri "gs://my-bucket/file.txt"
expunct redact --uri "s3://my-bucket/file.txt" --no-wait

JSON output (for scripting & AI agents)

expunct redact --text "My SSN is 123-45-6789" --json
expunct detect --text "Jane Doe" --json | jq '.findings[] | .entity_type'

---

🧱 How It Works

Your App / Logs / Files
           ↓
        Expunct
           ↓
   Clean, safe data
           ↓
AI / APIs / Analytics

Expunct acts as a privacy layer that removes sensitive data before it leaves your system.

---

📖 Commands

expunct redact

Redact PII from text, files, or URIs.

expunct redact --text "Call me at 555-1234"
expunct redact notes.txt
expunct redact scan.pdf -o redacted_scan.pdf
expunct redact --uri "gs://my-bucket/file.txt"
cat file.txt | expunct redact

Flag	Description
--text, -t	Inline text to redact
--uri, -u	Cloud URI to redact
--output, -o	Output file path (required for binary formats)
--language, -l	Language code (default: en)
--policy-id, -p	Policy ID to apply
--json	Raw JSON output
--wait/--no-wait	Wait for URI jobs (default: --wait)
--timeout	Wait timeout in seconds (default: 300)

expunct detect

Detect PII entities without redacting. Shows entity type, value, confidence, and location.

expunct detect --text "My name is Jane Doe and my SSN is 123-45-6789"
expunct detect document.txt
expunct detect --uri "gs://bucket/file.txt"
echo "[e-mail]" | expunct detect

expunct jobs

Manage redaction jobs.

expunct jobs list
expunct jobs list --status completed --page 2
expunct jobs get JOB_ID
expunct jobs download JOB_ID -o output.pdf
expunct jobs wait JOB_ID --timeout 600

expunct policies

Manage redaction policies.

expunct policies list
expunct policies create --name "strict" --confidence-threshold 0.9
expunct policies get POLICY_ID
expunct policies update POLICY_ID --name "updated-name"
expunct policies delete POLICY_ID --yes

expunct audit

View audit logs.

expunct audit list
expunct audit list --event-type redaction --page-size 50
expunct audit list --json

Document Intelligence commands (beta)

parse, extract, and safe-parse call the Expunct Document Intelligence API, which is currently in beta and feature-flag gated. Your tenant must be enabled before these commands will succeed; calls from disabled tenants return a 403. PDF and DOCX are the supported formats during beta.

expunct parse

Parse a PDF or DOCX into a canonical document structure (text, tables, headings, layout).

expunct parse report.pdf
expunct parse contract.docx --language en
expunct parse report.pdf --no-wait          # submit and return job ID immediately
expunct parse report.pdf --json             # raw JSON output

Flag	Description
--language, -l	Language code (default: en)
--wait/--no-wait	Wait for job to complete (default: --wait)
--timeout	Wait timeout in seconds (default: 300)
--json	Raw JSON output

expunct extract

Extract structured fields from a parsed document using a JSON Schema or built-in template.

# From an existing parse artifact (preferred — avoids re-parsing)
expunct extract art-abc123 --schema invoice_schema.json

# Parse and extract in one step
expunct extract --file report.pdf --template invoice

# Inline schema
expunct extract art-abc123 --schema-json '{"type":"object","properties":{"total":{"type":"number"}}}'

Flag	Description
PARSE_ARTIFACT_ID	Canonical document artifact ID from a completed parse job
--file, -f	PDF or DOCX to parse and extract in one step (mutually exclusive with artifact ID)
--schema, -s	Path to JSON Schema file defining fields to extract
--schema-json	Inline JSON Schema string
--template, -t	Built-in template ID (e.g. invoice)
--language, -l	Language code (default: en)
--wait/--no-wait	Wait for job to complete (default: --wait)
--timeout	Wait timeout in seconds (default: 300)
--json	Raw JSON output

expunct safe-parse

Parse a PDF or DOCX and sanitize PII in a single workflow. Produces sanitized canonical document, sanitized markdown, and sanitized chunk artifacts suitable for AI ingestion (RAG, prompts, embeddings).

expunct safe-parse contract.pdf
expunct safe-parse report.docx --policy-id strict
expunct safe-parse contract.pdf --no-wait        # submit and return job ID immediately
expunct safe-parse contract.pdf --json           # raw JSON output

Flag	Description
--policy-id, -p	Redaction policy ID applied during sanitization
--language, -l	Language code (default: en)
--wait/--no-wait	Wait for job to complete (default: --wait)
--timeout	Wait timeout in seconds (default: 300)
--json	Raw JSON output

expunct config

Manage CLI configuration.

expunct config set api_key YOUR_API_KEY
expunct config set base_url https://api.expunct.ai
expunct config get base_url
expunct config show
expunct config path

---

🔑 Authentication

Option A: Environment variable (recommended for CI/scripts)

export EXPUNCT_API_KEY=your_api_key

Option B: Config file

expunct config set api_key YOUR_API_KEY

Stored in ~/.expunct/config.json:

{
  "api_key": "your_api_key",
  "base_url": "https://api.expunct.ai",
  "tenant_id": "your-tenant-id"
}

Variable	Description
EXPUNCT_API_KEY	API key (overrides config file)
EXPUNCT_BASE_URL	API base URL (overrides config file)
EXPUNCT_TENANT_ID	Tenant ID (overrides config file)

---

⚙️ Output Modes

# Default: human-readable with rich formatting
expunct redact --text "My email is [e-mail]"

# JSON: machine-readable for piping and scripting
expunct redact --text "My email is [e-mail]" --json

---

🤖 Agent-Friendly Design

The CLI is designed to work well with AI agents and automation:

• Deterministic output with --json flag
• Stdin piping for chaining commands
• Non-interactive — no prompts in --json mode
• Exit codes — 0 for success, 1 for errors

# Pipe redacted text
expunct redact --text "My phone is 555-0100" --json | jq .redacted

# Chain with other tools
cat logs.txt | expunct redact | grep "ERROR"

# Use in scripts
expunct jobs list --json | jq '.jobs[].id'

---

🖥️ Platform Support

The CLI is pure Python and works on macOS, Linux, and Windows:

# All platforms via pip
pip install expunct-cli

# macOS via Homebrew (coming soon)
brew install expunct/tap/expunct-cli

---

🔒 Built for Developers

Expunct is built on top of proven detection tools like Microsoft Presidio, with added:

• policy control
• hosted API
• scalable processing
• multi-format support (text, PDF, DOCX, images, video, audio)

---

📚 Documentation

👉 https://docs.expunct.ai

🌐 Platform

👉 https://expunct.ai

---

💡 Roadmap

• Pseudonymization (reversible identity masking)
• Secret detection expansion
• Batch processing via CLI
• Directory processing
• Streaming mode
• Self-hosted / VPC deployment

---

🤝 Contributing

Contributions welcome. Feel free to open issues or PRs.

---

📄 License

MIT