Stays in your VPC
Philter deploys into your private subnet. No data leaves your perimeter; no third-party API path. HIPAA Safe Harbor, GDPR, and PCI-friendly out of the box.
Cloud Deployment
Deploy Philter in 5 minutes
Launch Philter — open source PII and PHI redaction — into your own AWS, Azure, or Google Cloud VPC. Billed through your existing cloud account. 30 days free.
$0.49/hour on AWS Marketplace
~$350/month if running 24/7
30-day free trial
Up to 18% off with annual contract
What you get
Marketplace deployments give you the full Philter feature set, running inside your VPC, billed through your existing cloud account.
30+ entity types
Names, ages, ID numbers, SSNs, credit cards, VINs, dates, IPs, addresses, and more — plus custom dictionaries and regex patterns for your domain-specific identifiers.
One unified bill
No procurement contract, no per-seat fees. Philter is metered through your existing AWS, Azure, or Google Cloud account — the same invoice as the rest of your infrastructure.
Healthcare & COVID lenses
Beyond the general-purpose NLP lens, Philter ships purpose-trained models for clinical text and COVID-era documents — higher accuracy on PHI without custom training.
API + SDKs + Office Add-Ins
Call the HTTP API directly, or use the open source SDKs for Java, .NET, and Go. Microsoft Word and Excel Add-Ins included.
Apache 2.0 source
Marketplace billing for the ops, Apache 2.0 for the trust. Audit the code on GitHub; no vendor black box.
Pick your cloud
All three marketplaces deploy the same Philter AMI/image into your account. Pick the one your team is already using.
Launch in 5 minutes
- Subscribe in your marketplace. Click any "Launch" button above. Marketplace flow asks you to accept terms, pick a region, and pick an instance type. Defaults are sensible.
- Launch the AMI/image into your VPC. Choose a subnet (private subnets are supported), security groups, and a key pair. For air-gapped deployments, attach to a subnet with no NAT.
- Wait ~2 minutes for the instance to boot. Philter starts as a systemd service on port 8080. Health check at
/api/status. - POST text to
/api/filter. From inside your VPC, the API accepts text and returns redacted text:$ curl http://<instance-ip>:8080/api/filter \ --data "His SSN was 123-45-6789." \ -H "Content-type: text/plain" His SSN was ***********. - Tune policies for your data. Per-entity strategies (redact, mask, encrypt, replace, drop) and confidence thresholds are configured via JSON. See the User’s Guide.
Marketplace or self-host?
Same software, two paths. Pick what fits how your team operates.
Choose marketplace when
- You want production-ready Philter in minutes, not days.
- Your procurement team prefers cloud-bill consolidation over a separate vendor contract.
- You want vendor-supported updates and patches.
- You’re evaluating Philter and want a 30-day trial without commitment.
- Per-hour pricing fits your usage pattern.
Choose self-host when
- You’re air-gapped or in a regulated environment that doesn’t allow marketplace AMIs.
- You need to build Philter into a custom container or Kubernetes deployment.
- You want to modify the source for your own redaction strategies.
- Your team is comfortable with Java build tools and operating JVM services.
- You want the open source tier with no marketplace metering.
Frequently asked questions
What does the 30-day free trial include?
Full access to all Philter features — the same software the paid tier runs. After 30 days the trial auto-converts to the paid hourly rate unless you cancel beforehand. Cancellation is one click in the marketplace; you keep what you’ve already deployed and stop being billed going forward.
How much will this actually cost me?
On AWS Marketplace, $0.49/hour across every supported instance type. Worked numbers:
- ~$350/month if running a single instance 24/7
- ~$290/month with the 365-day annual contract (~18% off)
- $0 for the first 30 days
Instance type matters for throughput, not cost. A t3.large is enough for most evaluation workloads; production redaction at scale typically runs on m5.xlarge or c5.xlarge. Pricing on Azure and Google Cloud follows similar per-hour models — check each listing for the current published rate.
Can I run in an air-gapped VPC?
Yes. After the AMI is in your account, you can attach the instance to a subnet with no NAT gateway and no internet egress. Philter does not phone home, does not download models at runtime, and does not require any outbound traffic to function. This is a common deployment pattern for healthcare and government workloads.
What instance type should I start with?
For evaluation: t3.large (2 vCPU, 8 GiB). For production: m5.xlarge (4 vCPU, 16 GiB) is a reasonable starting point; scale horizontally behind a load balancer when throughput demands it. GPU instances (p3.*) are available if you’re running the Healthcare lens at high volume.
Can I bring my own license (BYOL)?
BYOL is not currently offered on the marketplace listings. If your team has procurement requirements that the marketplace doesn’t fit, contact us — we can structure a direct engagement or private offer.
What support is included?
Email support from the Philterd team is included with marketplace deployments (typical response: within one business day). For named-contact SLAs, dedicated technical advisors, or hands-on engineering, see the consulting services or contact the team directly.
Ready to launch?
Pick your cloud, subscribe, and you’re running Philter in your VPC inside of 5 minutes. 30 days free; no procurement contract; cancel anytime from the marketplace.