Deployed on-premise, the model runs fully inside your network — no internet egress, nothing sent to the vendor. Summaries and risk signals appear in the console.
Schedule a Demo
Most AI-powered security tools route your intercepted data to a cloud endpoint — Anexet keeps every inference inside your own infrastructure.
Most AI-powered security tools route your intercepted data to a cloud inference endpoint — which means your sensitive communications, documents, and incident evidence leave the network boundary before any analysis happens. For banks, government agencies, defense contractors, and organizations operating under data residency requirements, that round-trip is not an option.
Anexet's AI Server module supports both on-premise and cloud models, configured independently. Run on-premise, the model operates entirely inside your own infrastructure — no internet egress, no vendor access, no data leaving your perimeter. Interceptions stay local. Summaries and statuses are generated on your own hardware, then written back into the Client Console.
The AI Server is also multimodal: analysis is not limited to text messages and documents, it extends to other file formats captured across Anexet's interception channels. Administrators can add custom context and instructions — sector terminology, data classification rules, internal policy — so the model's output reflects your organization rather than a generic baseline.
The result is a genuine on-premise AI capability: the same incident summarization, false-positive filtering, and automated status assignment you would expect from a cloud-connected tool, but entirely under your own control, on infrastructure that never leaves your network. When regulators or auditors ask where your incident data goes, the answer is: it does not leave.
Run on-premise, intercepted communications, documents, and other file formats are analyzed on your own server hardware — nothing is sent to the vendor or any external service.
The AI Server analyzes documents and other file formats alongside text communications, and takes admin-configured context and instructions into account to improve accuracy for your organization.
On-premise and cloud models are configured independently. For banks, government agencies, and any organization under data residency requirements, the on-premise mode keeps every inference — and every byte of source data — inside the network perimeter.
All capabilities below make up the AI Server module, running fully inside your Premium deployment.
Runs a language model inside your server perimeter; no internet connection required for inference; no data transmitted to the vendor or any cloud service. Cloud models are also supported for organizations that prefer that mode.
The model reads intercepted messages, emails, and documents attached to a flagged event and writes a concise summary directly in the incident viewer inside the Client Console.
The AI Server classifies incidents and sets their status automatically, removing routine triage from the security analyst's queue.
The model scores incidents and suppresses low-risk events so analysts focus only on genuinely suspicious activity.
When reviewing an intercepted document or conversation in the Client Console, a dedicated AI panel shows the model's assessment of the content alongside the raw material.
A dashboard within the Client Console shows how many incidents have been processed by the AI layer, status distribution, and filtering results.
The AI layer is multimodal: it analyzes not only text from intercepted emails and chats, but also documents and other file formats captured by Anexet's interception channels.
In the Administrator Console, security teams add custom instructions and context — sector, data classification rules, internal terminology — so the model's analysis reflects the organization's specific policies and improves accuracy, not a generic baseline.
The AI Server supports both on-premise and cloud models. Organizations that prefer a cloud model can connect one; on-premise and cloud modes are configured independently and can coexist.
The AI Server is a server-side component. Endpoint agents for interception are available for Windows, Linux, and macOS. Exact configuration is scoped together with your team during the demo.
Anexet agents on Windows, Linux, and macOS endpoints capture communications across email, messengers, web, cloud storage, USB, and print channels. All intercepted content — text, documents, and other file formats — is indexed and stored by the Data Processing Server inside your network. The AI Server then reads that data, applying any admin-configured context and instructions, without it ever leaving the perimeter.
Anexet is deployed entirely on-premise, and the AI Server itself can run in on-premise mode. The vendor has no access to your data. When the on-premise model processes an incident, it receives only content that is already inside your infrastructure. No cloud call is made, no data is sent to a third-party service, and no egress log will ever show traffic to an external AI endpoint — a decisive advantage for banks, government agencies, and any organization subject to data residency requirements.
Four steps from an intercepted event to an AI-assessed, analyst-ready incident record — entirely inside your network.
Anexet agents on Windows, Linux, and macOS endpoints capture communications across email, messengers, web, cloud storage, USB, and print channels; intercepted content — text, documents, and other file formats — is indexed by the Data Processing Server inside your network.
1
When the Security Policies module flags an incident — or on a configurable schedule — the AI Server reads the raw intercepted content associated with that event, applying any admin-configured context and instructions; in on-premise mode, all processing happens on your own server hardware.
2
The model produces a plain-language summary of the incident, assigns a risk status, and decides whether the event warrants analyst attention or can be suppressed as a false positive — across text, documents, and other formats.
3
The security analyst opens the incident in the Investigations or Security Policies module and sees the AI summary alongside the original intercepted content — and can accept, override, escalate, or dismiss it without the data ever leaving the network.
4
Anexet agents on Windows, Linux, and macOS endpoints capture communications across email, messengers, web, cloud storage, USB, and print channels; intercepted content — text, documents, and other file formats — is indexed by the Data Processing Server inside your network.
1
When the Security Policies module flags an incident — or on a configurable schedule — the AI Server reads the raw intercepted content associated with that event, applying any admin-configured context and instructions; in on-premise mode, all processing happens on your own server hardware.
2
The model produces a plain-language summary of the incident, assigns a risk status, and decides whether the event warrants analyst attention or can be suppressed as a false positive — across text, documents, and other formats.
3
The security analyst opens the incident in the Investigations or Security Policies module and sees the AI summary alongside the original intercepted content — and can accept, override, escalate, or dismiss it without the data ever leaving the network.
4
Common questions about Anexet's on-premise AI model and how it keeps incident data inside your perimeter.
On-premise AI security analysis means running a language model inside your own server infrastructure — without sending intercepted data to any external cloud or vendor endpoint — to automatically summarize security incidents, filter false positives, and assign risk statuses. Anexet's AI Server module delivers this capability on the Premium plan.
No. The default and recommended configuration uses a local model running entirely inside your network. No intercepted communications, documents, or incident evidence are transmitted to the vendor or any cloud service. Anexet is deployed on-premise; the vendor has no access to your data.
Yes. The AI Server module supports both on-premise and cloud model configurations. Organisations that permit external model calls can configure a cloud endpoint alongside the local model. Both modes are configured in the Administrator Console and operate independently.
The AI Server module, including local on-premise model support, is available exclusively on the Premium plan. Standard and Advanced plans do not include AI Server functionality. Contact sales for pricing — all plans are individual.
In the Administrator Console, security administrators supply additional context — sector-specific terminology, data classification categories, internal policy notes — that the model uses as a reference frame during analysis. This improves accuracy for your specific environment without requiring model retraining.
The AI Server assigns a status to each processed incident. Events classified as low-risk or false positives are filtered from the active queue, reducing analyst workload. The full incident record remains in the system; analysts can review suppressed events at any time through the Client Console search and Investigations module.
The AI Server is a server-side component that runs on your organization's server infrastructure (Linux-based server environments are supported). Endpoint agents — which perform interception — are available for Windows, Linux, and macOS. No mobile agents are available.
Yes. The AI Server is multimodal — it analyzes documents and other file formats captured across Anexet's interception channels, not just text communications, giving analysts a fuller assessment of each incident.
Run in on-premise mode, the AI Server keeps every model call, every piece of source data, and every generated summary inside your network perimeter — nothing crosses into the vendor's or a third party's infrastructure. This helps organizations comply with data residency and data sovereignty requirements common in banking, government, and other regulated sectors. Anexet is not certified under any specific regulation; it is designed to help you comply with your applicable requirements.
Tell us about your security environment and data sensitivity requirements — we will map the AI Server module to your infrastructure and arrange a demo or a free trial of Anexet Ultimate.
