Tracehound isolates runtime threats. Argos detects behavioral anomalies. Both are priced independently — choose what fits your security posture.
No request counting. No surprise bills. One price per logical service.
Developer / POC
Best for: Developers, Hobbyists, POC
Growing SaaS
Best for: Growing SaaS, Bootstrapped Startups
Cluster / High Scale
Best for: Scale-up, Fintech, High-Traffic E-commerce
Runtime Deviation Signals
Licensed and operated independently from Tracehound.
Standalone
With Tracehound Pro
Argos is a security signal layer that detects runtime deviation outside the request lifecycle.
Argos exists to surface security-relevant behavior that happens between requests.
"Argos detects when the runtime starts behaving differently — not why."
Argos samples the runtime and emits behavioral signals when something deviates from expected behavior. It observes:
Signal: "Did something behave differently than it should have?"
Argos never:
Signals are advisory and must be validated by external systems.
Argos is priced separately because it introduces a new security observation layer into your runtime. Unlike Tracehound, which preserves and isolates confirmed threats, Argos emits non-authoritative runtime deviation signals that must be verified by external detectors.
This expands the runtime trust boundary and adds operational responsibility, described in its own SLA.
Tracehound is not priced on traffic, usage, or attack volume — because none of those are the problem it solves.
Most security tools tolerate nondeterminism: Garbage-collected memory, Retry storms, Best-effort async pipelines. Tracehound does not.
Maintaining deterministic behavior under memory pressure, pool exhaustion, and partial failure requires explicit ownership models, bounded degradation paths, and strictly constrained concurrency. This engineering effort does not scale with usage. It scales with correctness guarantees.
In Tracehound, every failure mode is intentional: What degrades, What never degrades, What is preserved at all costs.
Designing and maintaining these semantics — without blocking the host application — is the core value of the system. This is not feature work. This is failure engineering.
Enterprise pricing exists because: Data shapes must remain stable, Evidence must remain verifiable, Behavior must remain explainable months or years later.
That requires versioned contracts and long-term support. Those guarantees have real cost.
The price difference reflects who is accountable when things go wrong.
If your system can tolerate undefined behavior during security incidents, you do not need Tracehound.
If it cannot, this is what precision costs.
Learn about per-service pricing, Kubernetes replicas, microservices, and more.
Read our FAQ →