Inventory with context
Devices, software and technical information help define the scope of the work.
The Myrmex product
From a connected environment to action that resolves the issue.
Myrmex brings AI agents for IT and security operations into one workspace. Connect your sources, investigate your environment and follow authorized actions with recorded results.
Map the connections
Topology and the technical conversation side by side, with the environment’s context.
Illustrative screens with demonstration data. Actions depend on the configured integrations, permissions and policies.
Context starts with what is in your environment.
Asset inventory, network topology and cloud sources help identify dependencies and guide diagnosis across connected systems.
Working contextServers
Networks
Cloud
Tools
Topology for investigation
Connections and subnets guide diagnostics, distinguishing observation from inference.
Organization that follows your team
Group assets by environment, business unit or client, with dedicated folders and instructions.
Inspect systems, packages and information from devices that responded to collection.
Asset inventory
Scan for vulnerabilities. Organize the fix.
Myrmex scans supported devices for vulnerabilities. It links findings to installed packages and versions, brings together risk signals, and hands remediation to AI with authorization and result verification.
Vulnerabilities found on the device
Illustrative screens with demonstration data. Actions depend on the configured integrations, permissions and policies.
Detect
Scanning and inventory identify vulnerabilities in the device’s packages.
Prioritize
CVSS, EPSS, KEV and environment evidence help order the work.
Remediate
AI receives the target and the fix to execute within the authorizations.
Verify
A new check shows what was resolved and what remains open.
One update can resolve several findings.
Myrmex groups CVEs by the update that fixes them. Your team can send the group for remediation, with the package, target version and related vulnerabilities in the context of the action.
Findings without an available update remain visible for investigation and risk treatment.
Explore plans with vulnerability scanningRelated CVEs
Package update
Fix for review
Risk needs to make sense before it becomes a priority.
Analysis brings together posture, audits, vulnerabilities, threat intelligence and alerts. CVSS, EPSS and asset evidence help prioritize investigation and remediation.
The score directs your attention. Evidence supports your decision.
Inspect controls, findings and coverage to understand what was assessed and what still needs verification.
Posture and auditing
Finding the issue is the start. Follow the fix.
In this example, an audit finds an inappropriate file permission on a server. Explore the plan, authorized execution and fresh verification: every stage has a result the team can inspect.
AI acts within the scope. Your team follows what changed.
Remediate the permission finding
The audit identifies an incorrect file permission on the server.
The same issue across several servers?
Group hardening controls and affected hosts into one change. Review the targets and impact; execution keeps steps and results for each device.
Available for controls and devices with supported fixes. Authentication and connectivity changes require attention to operational impact.
Selected hosts
Select the controls and targets.
One change for review
Review impact, authorization and the change window.
Results by device
Check execution and verification for each host.
When a fix affects production, the change needs the full picture.
Myrmex proposes infrastructure-as-code updates, prepares changes and organizes network execution. Permissions, review, maintenance windows and verification accompany the action.
Finding in the environment
Reviewable proposal
Authorized change
Recorded verification
The bucket finding leads to a Terraform proposal and a pull request for review. Opening the PR does not change the account.
Fix at the source
Recurring work can continue after the conversation ends.
Playbooks and runbooks can start on demand, on a schedule or through a webhook. The team reviews the workflow and authorizes its scope; actions beyond that scope require a new decision.
From description to procedure
Combine AI instructions, tool actions and scripts. Conditions and loops adapt the workflow to what each step finds.
On demand
On a schedule
By webhook
Check the context
Run permitted steps
Record the output
Anything requiring further authorization stays pending for a decision.
AI prepares a draft with a trigger, steps and actions that write to the environment. The team reviews it before enabling it.
Build the automation
The conversation continues with the environment beside it.
Mention devices and integrations, open evidence in tabs and keep the work in the selected context. Your team can follow the tools consulted and resume earlier investigations.
Resources in the conversation
Use mentions to identify the assets and sources that belong to the request.
Tabs and split view
Inspect a device, report and conversation without losing your place in the investigation.
Files and instructions
Add documents and guidance about the environment to inform the agents’ work.
Specialists on the same task
Centurion coordinates specialized agents to consult tools, investigate the environment and document the result.
“Investigate this asset and document what the next team needs to know.”
DeviceIntegrationDocumentEnvironment rules
Context follows the investigation and documentation.
Access is part of operations. It needs context, too.
An AI conversation does not remove responsibility for credentials, permissions and changes. Myrmex brings these controls together to define who can inspect and act in each environment.
Credentials outside the conversation
Use shared credentials or individual accounts with compatible connectors. Secrets are used by the integration and do not need to be included in the AI request.
Integration credentials
Roles and asset groups
Define capabilities and the set of resources each role can access.
Contexts and sub-organizations
Organize access by environment or client, preserving the operation’s scope.
Rules and approvals
Guide the work and apply authorization controls to changes.
Execution identity
Automations use a defined identity. Approval does not extend that identity’s technical permissions.
What you discover needs to stay useful to other people.
Technical reports, documentation and runbooks preserve consulted sources, decisions and open questions for other team members.
From technical detail to the conversation with decision-makers.
Scribe brings the observed environment into a technical document for review and operational continuity.
Environment documentation
Start with the tools your team already uses.
Connect the environment through agents and integrations. Reading and action capabilities vary by connector, configuration and granted permissions.
Endpoint agent
Information collection and local actions according to the agent’s configuration.
Environment connectors
Access to tools and devices using the integration’s credentials and capabilities.
Fortinet
MikroTik
Cisco
Palo Alto Networks
pfSense
Ubiquiti
Inspect devices and prepare actions with each vendor’s configuration in mind.
Explore the integration catalogBefore connecting your environment
Does Myrmex replace all the tools we already use?
The approach is to work with connected sources: monitoring, detection, ITSM, cloud and devices. The scenarios show Myrmex using these integrations to investigate and act in the environment.
Is everything AI proposes executed automatically?
No. A proposal, an automation draft and an approved change have different states. Execution must respect scope, permissions, policies and, where applicable, the defined window.
Can we start with read-only access?
Yes, with connectors that offer this mode. The team can assess context and diagnostics before enabling actions. Approval does not turn a read-only connection into a writable one.
How do we evaluate the features for our situation?
In a demo, we can walk through the sources relevant to your operation, a representative problem and the controls needed. Coverage depends on the connectors and environment configuration.
See the product working on the problems your team receives.
Bring an operational scenario. We will walk through context, investigation, proposal and evidence of the result.