Your Data Stays Home

Your data stays put.
Pipeline stays in control.

Sentra is a distributed data preprocessing platform that runs lightweight agents on your infrastructure. Orchestrate jobs, ensure compliance, and maintain data residency — all without moving sensitive data.

10B++
Data points processed
500++
Agents deployed
SOC 2+
Compliance certified
Zero data movement

Data never leaves
your infrastructure.

Sentra orchestrates preprocessing jobs across your infrastructure. Agents process data locally. The control plane logs what happened — it never sees the data itself.

Terminal — sentra agent

$ sentra job run --pipeline redact-pii

✓ Plugin verified (ed25519 signature match)

✓ Sandbox ready (mem: 512mb, cpu: 2, timeout: 300s)

✓ Processing 12,847 records...

✓ Done. 0 records containing PII — all clear.

$ _

Why choose Sentra

Data preprocessing, simplified

Build, run, and monitor data pipelines without the complexity. Focus on your data, not infrastructure.

Scale infinitely

Process terabytes of data across hundreds of agents. No infrastructure to manage.

Built-in compliance

SOC 2 compliant. Data never leaves your infrastructure. Full audit trails included.

Get started in minutes

Simple setup with our SDK. Deploy your first pipeline faster than you can read this.

Pay for what you use

Transparent pricing. No hidden fees. Start free, scale as you grow.

Request early access

Limited spots available for pilot program

Zero-Knowledge Architecture

Control plane sees outcomes. Not data.

Sentra orchestrates jobs, evaluates results, and tracks agent health. It knows what ran, whether it passed. It never receives the data processed.

Control Plane

A hosted backend that dispatches jobs, stores policy, tracks agent health, and records outcomes. Orchestrates everything — receives nothing sensitive.

  • Job orchestration & scheduling
  • Policy management & version control
  • Agent health monitoring
  • Audit logging & compliance

Agent

A Go binary that runs on your infrastructure. Downloads signed plugins, executes in sandbox, reports outcomes. Data stays where it lives.

  • Runs on-premise or in your cloud
  • Executes signed & verified plugins
  • Sandboxed execution environment
  • Reports outcomes, not data
Built for teams

Built for teams that own their data

Not a managed pipeline. Not a cloud ETL. Sentra is for teams where data residency is a hard constraint.

Data & ML engineers

Orchestrate preprocessing across distributed machines. Track what ran, on what data, whether it passed — without centralising data.

Privacy & security teams

Enforce redaction policy at the source. Every plugin signed. Every execution logged. No data leaves the processing environment.

Platform & infra teams

Deploy a single agent binary. Manage plugins centrally. Get consistent preprocessing behaviour across every environment.

How it's structured

Three components with distinct responsibilities. Each can be understood, audited, and replaced independently.

Agent

A Go binary that runs on your infrastructure. Downloads signed plugins, executes in sandbox, reports outcomes.

Control plane

Hosted backend that dispatches jobs, stores policy, tracks agent health, records outcomes. Receives nothing sensitive.

Plugins

Signed, sandboxed executables that do the work. Uploaded once, verified automatically on every agent.

What Sentra guarantees

Every design decision traces back to one requirement: prove what happened to your data, without centralising the data itself.

Cryptographic verification

Every plugin signed with Ed25519 key. Agent verifies at startup and before execution.

Sandboxed execution

Each plugin runs with enforced memory, CPU, and wall-clock limits.

Explicit outcomes

Every job produces a recorded outcome. Pipeline stops on failure.

Early Access Program

Join our pilot program

We're scaling up gradually to ensure the best experience. Book a call to get early access.