Salus
Winter 2026 NewGuardrails to validate your agent's actions before they execute
Your agent processed a refund without looking up the order ID, costing you thousands. You only found out three hours later from a support ticket. Evals, output scoring, and observability can reduce the likelihood of mistakes like these occurring - but there's no solution that inspects and prevents an action as it’s about to execute. Salus does that. We’ve built an API that wraps around your agent and checks its actions at run time, blocking incorrect ones and providing immediate feedback to guide retries. Kevin and Vedant were roommates at Stanford, where they both studied computer science.
AI Investor Summary
Salus provides critical runtime guardrails for AI agents, preventing costly errors before they occur. With a strong technical team from Google and Meta, they are well-positioned to capture a massive and rapidly growing market for AI agent safety and reliability. Their API-first approach offers a technically differentiated solution to a pressing industry need.
Key Highlights
- ● Addresses a critical and growing pain point in AI agent deployment.
- ● Founders possess strong technical backgrounds from top institutions and leading tech companies.
- ● Significant market opportunity with strong tailwinds.
- ● Innovative product approach with runtime validation.
Risk Factors
- ● Execution risk in building a robust and scalable validation engine.
- ● Competition from broader AI observability platforms or internal solutions.
- ● Customer adoption and integration challenges for a new infrastructure component.
- ● Demonstrating clear ROI and cost savings to potential customers.
Founders
Kevin Pan is the co-founder of Salus, a Y Combinator startup focused on AI-powered solutions. His background appears to be in engineering and technology, with a focus on building innovative products. He has a strong academic foundation and has been involved in the startup ecosystem.
Vedant Singh is the co-founder of Salus, a Y Combinator startup focused on AI-powered security for the metaverse. His background includes significant experience in software engineering and a strong academic foundation in computer science.
Score Breakdown
Strong technical team with excellent academic pedigrees from Stanford and Berkeley. Both founders have prior experience as Software Engineers at top-tier tech giants like Google and Meta, indicating a high level of engineering talent and ability to build complex systems. This background suggests a strong founder-market fit for a technical infrastructure product. [Boost +1: Founder from Google; Founder from Google]
The market for AI agent safety and reliability is exploding. As AI agents become more autonomous and integrated into business processes, the need for guardrails to prevent costly errors is paramount. The TAM is massive, encompassing virtually any business deploying AI agents. Regulatory tailwinds are likely to emerge, further driving demand. Timing is excellent as this problem is becoming acutely felt. [Boost +0.5: Hot sector: ai]
The core concept of runtime validation for AI agent actions is technically differentiated and addresses a critical gap in the current AI observability and evaluation landscape. The API-first approach suggests platform potential. Defensibility will come from the sophistication of the validation engine and the breadth of integrations. UX quality is yet to be fully demonstrated at scale, but the core value proposition is clear.
Early stage with limited public traction data. The $3.7M funding round is a positive signal of investor interest, but specific revenue or user growth figures are not readily available. Partnerships and press coverage are nascent, indicating the company is still in its early go-to-market phase. [Boost +2: Tier-1 VC: accel; Tier-1 VC: accel]
News
Salus, founded in 2026, is developing a runtime security platform for validating and blocking agent actions, having raised $500K in seed funding.
Salus has raised $3.7M and offers a platform that validates AI agent actions in real-time, preventing incorrect actions and providing immediate feedback.
Salus provides runtime guardrails that validate AI agent actions before execution to prevent costly mistakes, hallucinations, and policy violations.
Salus has raised $3.7M, offering a platform that validates AI agent actions in real-time with features like policy configuration, evidence grounding, and runtime guardrails.
Salus offers AI agent runtime guardrails that validate actions before execution, focusing on autonomous systems, agent frameworks, and monitoring tools.
Salus is developing real-time, ML-driven guardrails that validate AI agent actions before execution, using hybrid LLM-based evaluators and evidence grounding to prevent unsafe or hallucinated outputs.
Salus has launched a new platform that provides runtime guardrails to validate AI agent actions before they are executed, aiming to prevent costly mistakes, hallucinations, and policy violations.
Salus, a Y Combinator Winter 2026 startup, offers runtime guardrails for AI agents, addressing a growing need in the market, though the founding team's experience and competitive landscape are noted as areas for development.
Salus, a San Francisco-based startup founded in 2026, has raised $500K in seed funding for its runtime security platform that validates and blocks agent actions.
Salus has announced the enterprise deployment of its Medication Administration Protection System (MAPS) within Houston Methodist, a healthcare technology company focused on medication safety.
Salus, a San Francisco-based startup founded in 2026, has raised $500K in seed funding and develops a runtime security platform for validating and blocking AI agent actions.
Salus offers an API that wraps around AI agents to validate their actions at runtime, blocking incorrect ones and providing immediate feedback for retries.
Salus provides runtime guardrails to validate AI agent actions before execution, offering self-repair capabilities and full visibility into agent interactions.
Salus, a healthcare technology company focused on medication safety, announced the enterprise deployment of its Medication Administration Protection System (MAPS) within Houston Methodist, a large academic health system.
Quick Info
- Batch
- Winter 2026
- Team Size
- 2
- Location
- San Francisco, CA, USA
- Founders
- 2
- Scraped
- 4/10/2026