Specialist agents that work within the limits you set
Ciyora Neural Mesh is a set of AI agents, each focused on one kind of work: triaging alerts, reviewing identities, summarising cloud findings, checking endpoint health, chasing compliance evidence and writing posture summaries. You choose how much each one may do on its own.
What Ciyora Neural Mesh does
Six specialist agents
Each agent is focused on one kind of work and uses only the tools that work needs. Every run keeps its reasoning steps, so you can see how an agent reached its result.
- Alert triage: groups and prioritises open alerts, enriches them and proposes a verdict
- Identity review: risky identities, stale access and separation-of-duties conflicts
- Cloud posture: misconfigurations and attack paths, summarised
- Endpoint health: agent health, versions and endpoint alerts
- Compliance audit: controls and evidence against your frameworks
- Security posture assessment: one view across detection, identity, cloud and endpoints
Autonomy is a setting, not a surprise
Each agent works under the autonomy level set for its module: off, suggest, ask first, act and report, or autonomous. An agent that wants to change something at a lower level proposes the action, and proposals wait for a person to approve or reject them.
- Enable or disable each agent, with confirmation
- Run an agent now with a specific task
- Run history with the reasoning steps of each run
- Pending actions approved or rejected by a person, and recorded
Investigations and playbooks
AI playbooks are ordered investigation recipes your team can read and edit. Investigations collect status, evidence, notes and a timeline, so an agent’s work and a person’s work sit in the same place.
- Built-in and custom playbooks by category
- Deep investigations that run in the background and report back
- A digest of critical alerts and a morning briefing for the team
Digital twin
Model an attack scenario against your real assets and alerts before it happens. A simulation shows the assets that would be affected, the alerts it would raise, the attack path and the controls that would stop it.
Your model, your data
Agents run on the language model your tenant configures: Anthropic, OpenAI, Amazon Bedrock or a self-hosted Ollama model. They read data only from your tenant.
How Ciyora Neural Mesh works with the rest of the platform
Every module shares one tenant, one identity model, one audit chain and one console, so a finding in one place is context everywhere else.
Ciyora Neural Mesh is included with every paid module. See pricing.
See Ciyora Neural Mesh in a demo
A demo walks through the module on a sample tenant and covers how it would connect to your identity provider, cloud accounts and sites.