Salus

Name: Salus
Brand: Salus

AI Native

Runtime guardrails to validate your agent's actions before they execute

Est 2026|Raised $500K

LinkedIn 115|Website|FTEs 2|HQ San Francisco, CA

01Competitive Landscape

| Aspect | Salus (official product page) | Invariant Guardrails | Humanloop | |--------|-------------------------------|----------------------|----------| | Starting Price | Not public (demo/sales-led) | Not public on referenced launch page | Free trial tier shown; enterprise is sales-led | | Appropriate Use Cases | Teams seeking commit-time validation of tool actions (payments, records, bookings) | Teams wanting contextual guardrailing for AI applications | Teams needing LLM evaluation, logging, and enterprise workflow support to ship LLM features | | Key Differentiator | Architectural placement as a runtime proxy that enforces policies on execution-path calls and records decision evidence (vendor claim) | Contextual security/guardrail framing for application inputs/outputs | Focus on evaluation lifecycle, dataset management, and CI/CD-style eval integration | | Primary Limitations | Limited public proof: pricing, certifications, and customer stories are not transparent | Pricing and production references not obvious from the referenced public page | Not primarily designed as a runtime block/allow gate for every tool call; can be adapted but is not the core product |

Narrative analysis

Direct competitors and key differentiators
- Salus emphasizes commit-time judgment for tool calls (block/clarify/repair/escalate prior to execution), which is an architectural approach focused on preventing harmful or incorrect actions at the point of execution.
- Many guardrails products focus on input/output moderation, prompt injection detection, or evaluation and monitoring, which may not enforce a hard gate on the action path.
Feature comparison: areas where Salus may excel vs lag
- Potential strengths
  - Placement in the execution path can provide consistent enforcement across agent frameworks and tool chains.
  - A shadow replay workflow (vendor-described) for importing prior calls and testing policies before enforcement could reduce deployment risk if implemented and maintained correctly.
- Potential weaknesses
  - If the runtime proxy becomes a central choke point, reliability and high-availability requirements increase.
  - Policy authoring and evidence requirements may be operationally heavy and could slow iteration if not well supported by tooling and integrations.
Pricing comparison
- Salus: unknown.
- Humanloop: publishes a free trial tier and positions enterprise offerings via sales; direct price comparison is not possible without Salus pricing information.
Market share and adoption trends
- Not enough public data to quantify adoption for Salus specifically.

01Competitive Landscape

Aspect	Salus (official product page)	Invariant Guardrails	Humanloop
Starting Price	Not public (demo/sales-led)	Not public on referenced launch page	Free trial tier shown; enterprise is sales-led
Appropriate Use Cases	Teams seeking commit-time validation of tool actions (payments, records, bookings)	Teams wanting contextual guardrailing for AI applications	Teams needing LLM evaluation, logging, and enterprise workflow support to ship LLM features
Key Differentiator	Architectural placement as a runtime proxy that enforces policies on execution-path calls and records decision evidence (vendor claim)	Contextual security/guardrail framing for application inputs/outputs	Focus on evaluation lifecycle, dataset management, and CI/CD-style eval integration
Primary Limitations	Limited public proof: pricing, certifications, and customer stories are not transparent	Pricing and production references not obvious from the referenced public page	Not primarily designed as a runtime block/allow gate for every tool call; can be adapted but is not the core product

Narrative analysis

Direct competitors and key differentiators
- Salus emphasizes commit-time judgment for tool calls (block/clarify/repair/escalate prior to execution), which is an architectural approach focused on preventing harmful or incorrect actions at the point of execution.
- Many guardrails products focus on input/output moderation, prompt injection detection, or evaluation and monitoring, which may not enforce a hard gate on the action path.
Feature comparison: areas where Salus may excel vs lag
- Potential strengths
  - Placement in the execution path can provide consistent enforcement across agent frameworks and tool chains.
  - A shadow replay workflow (vendor-described) for importing prior calls and testing policies before enforcement could reduce deployment risk if implemented and maintained correctly.
- Potential weaknesses
  - If the runtime proxy becomes a central choke point, reliability and high-availability requirements increase.
  - Policy authoring and evidence requirements may be operationally heavy and could slow iteration if not well supported by tooling and integrations.
Pricing comparison
- Salus: unknown.
- Humanloop: publishes a free trial tier and positions enterprise offerings via sales; direct price comparison is not possible without Salus pricing information.
Market share and adoption trends
- Not enough public data to quantify adoption for Salus specifically.

Salus

01Competitive Landscape

Narrative analysis

02Market Map

01Competitive Landscape

Narrative analysis

02Market Map