Licensing & Pricing

Plans for all sizes

Frictionless whole-system visibility across all major languages

Free

Free Forever

Perfect for early startups

  • Deployment Self-Hosted
  • License Type Perpetual
  • CPU Profiling
    Full Access
  • GPU Profiling
  • Profile-Guided Optimization
  • Recommendations and Alerts
  • Core Limit Limited to 5 concurrent machines
  • Support Discord Community
  • Fully Managed Deployment
  • Capacity Planning
  • Air Gap Installation Support
  • Training & Consulting

Standard

$1

/core/month

+ $0.25 /TFLOP/month for GPUs*

Volume discounts available

*For NVIDIA data center models (A100, A800, H100, H200, and later)

  • Deployment Self-Hosted
  • License Type Subscription
  • CPU Profiling
    Full Access
  • GPU Profiling
    Full Access
  • Profile-Guided Optimization

    Coming Soon

  • Recommendations and Alerts

    Coming Soon

  • Core Limit No Limit
  • Support Email & Chat
  • Fully Managed Deployment
    Additional fee
  • Capacity Planning
  • Air Gap Installation Support
  • Training & Consulting

Enterprise

Custom

[email protected] For customers profiling 30k+ cores

  • Deployment Self-Hosted
  • License Type Custom
  • CPU Profiling
    Full Access
  • GPU Profiling
    Full Access
  • Profile-Guided Optimization

    Coming Soon

  • Recommendations and Alerts

    Coming Soon

  • Core Limit No Limit
  • Support Priority Support
  • Fully Managed Deployment
    Additional fee
  • Capacity Planning
  • Air Gap Installation Support
  • Training & Consulting

Features

What makes zymtrace unique

zymtrace is a platform for monitoring and managing your code.

CPU Profiling

Frictionless whole-system visibility across all major languages. Identify bottlenecks and optimize performance with detailed insights at scale.

Supported Languages: Python, Java (and all JVM languages), Go, Rust, C/C++, Node.js/V8, Ruby, .NET, PHP, Perl, Zig, Scala, and more.

GPU Profiling

Identify performance bottlenecks in CUDA kernels, optimize inference batch size, and eliminate idle GPU cycles —with zero friction.

Supports: CUDA, PyTorch, JAX, and other major GPU frameworks.

Read the GPU profiler use-case ->

Profile-Guided Optimization

Coming soon

Imagine a docker container running FOSS software like Nginx. With the same software version, source code, and configuration, zymtrace’s profiling-guided techniques deliver up to 14% faster performance—no code changes required.

Recommendations and Alerts

Coming soon

zymtrace Curated insights. Most profilers throw flamegraphs at you and expect you to decode them. zymtrace’s “Efficiency IQ” tells you exactly what’s happening and shows you precisely what to do about it.

Deployment Options

All instances are self-hosted by default, giving you complete control over your data and infrastructure.

Standard and Enterprise customers can opt for our Fully Managed deployment service where we handle all aspects of installation, maintenance, and updates.

Enterprise Services

Enterprise customers benefit from our full range of professional services:

  • Comprehensive capacity planning
  • Air gap installation support
  • Dedicated training sessions
  • Expert consulting services

Need Integration Support?

For workloads exceeding 500,000 cores or enterprise-scale deployments, our team is available to assist with tailored migrations and integrations, provided they align with our roadmap.

Get started