Action validation; Policy checks; Audit evidence
Salus checks proposed AI-agent actions before a backend executes them. It compares consequential tool calls with the relevant policy and evidence from the current session, then allows, blocks or returns a correction for the agent to address. The official workflow also records the decision and checks completion by the downstream provider. This is an authorization and control layer around an existing agent stack. It is especially relevant to actions that affect records, payments or external systems, but it does not make every permitted action correct automatically. Published plans scale with registered agents and authorization checks rather than user seats, with both shadow and enforcement modes described.
Salus is best described as AI Security And Governance Tool for security teams, ai platform teams. The practical workflow centers on pre-execution action checks, policy and session-evidence checks, repair feedback, decision records, shadow and enforcement modes. Users normally bring their own task context into the product and review the resulting output before relying on it.
Useful use cases include action validation, policy checks, audit evidence. Category placement is kept to AI Security And Governance because the tool should be listed where people would actually compare it. Supported access is recorded as the access model described by the product, and integrations are limited to connections that could be verified from the available source material.
The developer is recorded as Salus. Pricing is listed conservatively as Paid plans based on agents and authorization checks. Free-plan status is recorded as Free access terms are not clearly established in the reviewed public material. Before using Salus for production work, check the current plan page, account limits and any commercial-use terms that apply to the files, data, media or decisions involved.
Run a small real task first and compare the result with the original material. For generated text, media, code, analysis or operational actions, review factual claims, permissions and handoff steps before publishing or applying the output. This keeps the listing useful without adding unsupported benchmarks, invented model names or broad legal promises.
