HyperProbe

AI-powered debugging tool that places virtual breakpoints in production to capture live variable state without redeployment.

Visit website

Devtool AI Hosting

Overview

HyperProbe is a production debugging platform for engineering teams investigating incidents in running services. Its SDK and read-only probes collect variable snapshots, logs and call stacks so developers can inspect runtime behaviour without repeatedly adding logging statements and redeploying.

The product connects to coding agents through MCP and provides a VS Code extension. Supported runtimes include JavaScript, TypeScript, Java, Kotlin, Python and Ruby; Go is listed on the roadmap.

Managed cloud is available, with self-hosting options on qualifying paid plans. Teams should evaluate capture policies, data redaction and access controls before deploying it to production. Performance and incident-resolution claims are vendor-reported and have not been independently benchmarked by Newtools.

Key features

Virtual breakpoints in production
Read-only, non-blocking probes
AI-powered root cause analysis
Incident workflow integrations, including PagerDuty and Datadog
MCP server for coding agents (Cursor, Claude Code)
Automatic code line identification
Probe lifetime, rate and hit controls
Immutable audit trail
PII redaction
Self-hosted or managed cloud deployment

Pros

Captures exact variable state without redeployment
Free plan available for one service
Works with silent failures and race conditions
Non-blocking and asynchronous
Integrates with existing incident workflows
Unlimited probes and captures on all plans
Supports multiple programming languages
Can run entirely within your infrastructure

Cons

Limited to 6 supported runtimes (Go on roadmap)
Requires SDK installation in your services
Free plan limited to 1 service
Professional plan requires 3-service minimum
Free plan uses managed cloud and retains capture history for 7 days

Use this if

You maintain services in a supported runtime and need runtime snapshots or logs to investigate production incidents. Evaluate its SDK and capture policies with your security requirements.

Skip this if

You only run applications in languages not yet supported (Go, Rust, etc.), or you have no production debugging needs.

Best for

Teams with senior engineers stuck on on-call rotationsProduction debugging without redeploymentFinding silent failures and race conditionsReducing incident resolution timeCapturing data that logs don't haveDistributed system troubleshooting