Built for the standard justice requires
A decade inside courts before putting agents in them. Security your IT team can verify, accessibility for everyone the system touches, and commitments we publish and keep.
Three things we don’t do
We don’t ask agencies to rip out what they have
Justice agencies have invested in systems that still do their jobs. eCourt AI connects to what exists, adds what’s missing, and lets your agency modernize one step at a time, on your timeline.
We don’t hand anyone a black box
We work alongside agencies one workflow at a time, with humans in the loop and results you can see and verify. Every agent action is attributed, timestamped, reviewable, and exportable.
We don’t build AI that replaces the people who serve communities
The agents give professionals their time back. The judgment, the discretion, the humanity: those stay where they belong. Each staff member staffs agents; agents never staff themselves.
Security & compliance
CJIS compliant
Built to the criminal justice information standard your agency already answers to.
AWS GovCloud
Justice data lives in the government cloud, inside the boundary your auditors expect.
Every action audited
Complete audit trails for staff and agents alike: who did what, when, and on whose approval.
No provider logging
Open-source models log to no one; API models run through AWS Bedrock, which does not log to the provider.
Consent rules enforced
Your jurisdiction’s consent and communication rules, including TCPA, enforced by the platform and not by memory.
Data portability
Your data is yours: exportable on request, in open formats, with no exit fees.
IT and security, answered
The questions a security review asks first, answered in writing. What this page doesn’t cover, ask in the first call and we answer with specifics.
- Where does justice data live?
- In AWS GovCloud, the U.S. government cloud, inside the boundary your auditors already expect. Which boundary a deployment runs in follows your jurisdiction's residency requirement, and it is stated in writing before you sign. The public website you are reading now is separate from the operating system agencies run.
- What does the model provider see?
- Open-source models run privately, with no third-party logging at all. API models run through AWS Bedrock, which does not log prompts or outputs to the model provider and never uses court data for training. The exact model list for a deployment is stated in writing before you sign.
- Is court data used to train AI models?
- No. AWS Bedrock never uses court data for training, and self-hosted open-source models train on nothing they process.
- Who can see what an agent did?
- Your agency. Every action is attributed to the agent that took it and the human who approved it, timestamped, reviewable, and exportable. Staff and agents share one audit standard.
- Can we export our data and leave?
- Yes. Exportable on request, in open formats, with no exit fees. We keep your business by being worth keeping.
- How do local rules stay local?
- Governance and compliance are configured to your jurisdiction, at every level that sets rules, down to an individual court’s own practice. One system runs an entire jurisdiction, with every local court on its own terms.
- How do staff accounts work?
- Staff sign in through your agency’s single sign-on (SSO), and role-based access control (RBAC) covers agents the same way it covers people. Nothing about an agent bypasses the permissions a person would need.
- What about consent rules for messages?
- Consent and communication rules are enforced by the platform, not by memory, configured to the rules your jurisdiction applies. In the United States that means TCPA.
- What happens when an agent is wrong?
- It is caught on the record, corrected once by a person, and the correction lands everywhere the original went. The full correction workflow is on the governance and controls page.
Accessible to everyone, in every language
Everything we build meets WCAG 2.2 AA and speaks 100+ languages: the staff interfaces, the public resources, the lobby tablets, the voice lines. The justice system serves everyone, so its technology has to.
A decade of earning it
We spent ten years working inside courts and justice agencies before putting agents in them. That is where the workflows, the guardrails, and the patience come from, and it is why we treat "how do we use AI responsibly" as the only question worth answering well.
Reporting a security issue
Good-faith security research protects the people this system serves, and we welcome it. If you find an issue on this site or any eCourt AI surface, email info@ecourt.ai with “Security” in the subject line. We read those first and reply within one business day. Machine-readable contact details are published at /.well-known/security.txt.
Put your IT and security team in the room
This page is written to be checked. Bring the checklist to the first call and we answer CJIS and AWS GovCloud, model governance, and audit questions with specifics.
Not sure where to start?Get an AI readiness plan