Assail home

DEPLOYMENT TYPE / BRING YOUR OWN CLOUD

Your cloud. Ares on the inside.

Your cloud. Ares on the inside.

Deploy Ares in a cloud account or project your organization operates. Plan private application access, choose where inference runs, and apply releases through your own change process with a supported AWS or Google Cloud configuration.

CUSTOMER CLOUD

Infrastructure · Inference · Updates

PLATFORM LOCATION ≠ MODEL LOCATION

Customer-hosted inference keeps prompts in your environment. Assail-hosted inference sends prompts to that service. Plan both boundaries.

DEPLOYMENT TYPE / BRING YOUR OWN CLOUD

Your cloud. Ares on the inside.

Deploy Ares in a cloud account or project your organization operates. Plan private application access, choose where inference runs, and apply releases through your own change process with a supported AWS or Google Cloud configuration.

CUSTOMER CLOUD

Infrastructure · Inference · Updates

PLATFORM LOCATION ≠ MODEL LOCATION

Customer-hosted inference keeps prompts in your environment. Assail-hosted inference sends prompts to that service. Plan both boundaries.

ARES DEPLOY / BRING YOUR OWN CLOUD

Your cloud. Ares on the inside.

Place Ares within the boundary you operate.

Deploy Ares in a supported AWS account or Google Cloud project. Plan private target routes, permissions, storage, and compute around your environment.

Choose model hosting separately. Review external data flows and capacity, then govern backups, recovery, and releases through your own change process.

Explore Ares

01

ARES DEPLOY / BRING YOUR OWN CLOUD

Your cloud. Ares on the inside.

Place Ares within the boundary you operate.

Deploy Ares in a supported AWS account or Google Cloud project. Plan private target routes, permissions, storage, and compute around your environment.

Choose model hosting separately. Review external data flows and capacity, then govern backups, recovery, and releases through your own change process.

Explore Ares

01

ARES DEPLOY / BRING YOUR OWN CLOUD

Your cloud. Ares on the inside.

Place Ares within the boundary you operate.

Deploy Ares in a supported AWS account or Google Cloud project. Plan private target routes, permissions, storage, and compute around your environment.

Choose model hosting separately. Review external data flows and capacity, then govern backups, recovery, and releases through your own change process.

Explore Ares

01

THE ARES METHOD

Workflow

Workflow

Plan the boundary, validate the deployment, and operate on your terms.

01

Define the boundary

Identify the account or project, authorized targets, data requirements, and model-hosting choice.

01

Define the boundary

Identify the account or project, authorized targets, data requirements, and model-hosting choice.

02

Prepare your cloud

Plan network routes, permissions, capacity, policies, and the services required by the supported configuration.

03

Install Ares

Use the supported deployment process. Establish administrator access and validate the inference path.

04

Validate target reach

Confirm scope, network access, and credentials. Run an authorized assessment and review its evidence.

05

Operate and update

Establish backup and recovery procedures. Review releases and have an operator apply approved changes.

Platform placement, inference location, and external dependencies are separate decisions.

THE ARES METHOD

Workflow

Plan the boundary, validate the deployment, and operate on your terms.

01

Define the boundary

Identify the account or project, authorized targets, data requirements, and model-hosting choice.

02

Prepare your cloud

Plan network routes, permissions, capacity, policies, and the services required by the supported configuration.

03

Install Ares

Use the supported deployment process. Establish administrator access and validate the inference path.

04

Validate target reach

Confirm scope, network access, and credentials. Run an authorized assessment and review its evidence.

05

Operate and update

Establish backup and recovery procedures. Review releases and have an operator apply approved changes.

Platform placement, inference location, and external dependencies are separate decisions.

Illustrative Ares product interface showing target selection and operation scheduling

CUSTOMER CLOUD · INFERENCE CHOICE · CONTROLLED RELEASES

Capabilities

Reach private APIs and internal applications through approved network paths.

Bring infrastructure placement, model hosting, data flows, and releases into one plan. Confirm the supported provider configuration and required workflows with Assail.

Read more

CUSTOMER CLOUD · INFERENCE CHOICE · CONTROLLED RELEASES

Capabilities

Reach private APIs and internal applications through approved network paths.

Bring infrastructure placement, model hosting, data flows, and releases into one plan. Confirm the supported provider configuration and required workflows with Assail.

Read more

CUSTOMER CLOUD · INFERENCE CHOICE · CONTROLLED RELEASES

Capabilities

Reach private APIs and internal applications through approved network paths.

Bring infrastructure placement, model hosting, data flows, and releases into one plan. Confirm the supported provider configuration and required workflows with Assail.

Read more

01 / CUSTOMER CLOUD

Bring the platform into your account.

Run the deployment on resources in your own cloud account or project. Govern network paths, infrastructure policies, permissions, and resource capacity as part of the deployment design.

02 / PRIVATE REACH

Start closer to the applications.

Place Ares within reach of authorized private APIs and internal applications. Routing, segmentation, credentials, and supported campaign-agent placement determine the functionality an operation can examine.

03 / MODEL HOSTING

Choose where the decisions happen.

Select customer-hosted or Assail-hosted inference during planning. Local processing requires the appropriate model configuration and compute; application placement alone does not keep prompts inside your cloud.

04 / FINDINGS & EVIDENCE

Define the full data boundary.

Use the storage and database resources configured for your customer environment. Plan access, retention, backups, and recovery, and include enabled integrations and exports in the data-flow review.

05 / CONTROLLED UPDATES

Keep deployment actions in your change process.

Assail provides the product release. Your team reviews and approves the version, then an operator applies it through the supported process with signature and verification checks.

ILLUSTRATIVE ARCHITECTURE

Decide what runs inside your boundary.

Decide what runs inside your boundary.

This conceptual customer-hosted inference configuration is not a fixed provider topology. It separates platform placement from model placement so both decisions remain visible.

YOUR CLOUD ACCOUNT OR PROJECT

ARES PLATFORM

Application services and the configured storage and database resources for findings and evidence.

CUSTOMER-HOSTED MODEL

Supported model-serving configuration and compute inside the customer environment.

AUTHORIZED TARGETS

Private applications and APIs reached through approved routes, identities, and agents where supported.

Customer-hosted inference keeps assessment prompts within this planned environment. Confirm the complete service configuration, capacity, and data flows.

REVIEW EVERY APPROVED EXTERNAL CONNECTION

Assail-hosted inference sends prompts to that service. Entitlement and update modes, exports, and integrations can also introduce external dependencies. BYOC is not a blanket zero-egress guarantee.

MODEL-HOSTING CHOICES

The platform and the model are separate decisions.

The platform and the model are separate decisions.

Choose inference placement during deployment planning so compute, networking, and data handling align from the start.

Configuration

Prompt-processing location

Planning requirement

Customer-hosted inference

Within your environment using the supported customer-hosted model configuration.

Confirm the provider configuration, required model compute, capacity, and network paths.

Assail-hosted inference

Prompts are sent to the configured Assail-hosted service.

Review the external dependency and its data flow against your requirements.

Application state, findings, and evidence use resources configured for the customer deployment. Define access, retention, backup, and restore requirements. Review exports and integrations alongside model traffic.

PROVIDER CONFIGURATION

Choose the architecture for the cloud you use.

Choose the architecture for the cloud you use.

Work with Assail on the supported provider configuration rather than assuming identical infrastructure, protections, or feature availability.

Environment

What to define

AWS

Customer account, landing zone, target routes, permissions, storage, model capacity, and release management.

Google Cloud

Customer project, approved application services, model compute, workload capacity, and provider-specific infrastructure and protection mechanisms.

Government requirements

Review AWS GovCloud or Google Assured Workloads requirements with Assail Federal, including model location, dependencies, and the applicable service evidence.

Confirm the offered release and required assessment lanes before deployment. Provider infrastructure does not establish Ares service certification, and BYOC does not imply SaaS feature parity.

OPERATOR-CONTROLLED RELEASES

Assail supplies the release. Your team applies it.

Assail supplies the release. Your team applies it.

Review release notes, approve the version, and have an operator apply it through the supported process. Coordinate the change with maintenance, validation, and recovery requirements.

01 / REVIEW

Evaluate the available release and its notes against your operating requirements.

02 / APPROVE & VERIFY

Authorize the version and use signed release and verification checks to establish the expected software.

03 / APPLY & VALIDATE

An operator performs the update through the supported process and checks the resulting deployment.

Plan recovery procedures explicitly. No zero-downtime update, automatic rollback, or fixed recovery-time guarantee is implied.

OPERATING RESPONSIBILITIES

Agree on ownership before installation.

Agree on ownership before installation.

A customer cloud deployment brings infrastructure responsibilities alongside product capability. Define the people, procedures, and capacity needed to operate it.

Area

What to define together

Cloud infrastructure

Account or project, approved regions, resource capacity, and infrastructure permissions.

Network access

Target routes, segmentation, required outbound dependencies, and any supported agent placement.

Model hosting

Customer-hosted or Assail-hosted inference, required compute, and resulting data flows.

Data operations

Access to findings and evidence, retention, backups, and restore procedures.

Releases

Approval authority, update operator, maintenance windows, and recovery process.

Support

Diagnostic information your team can provide and the agreed support procedure.

FROM DESIGN TO FIRST OPERATION

Validate the deployment and the assessment path.

Validate the deployment and the assessment path.

Start with the boundary, then confirm that the selected platform, inference path, and target access work together.

Step

Action

01 / Define

Identify the account or project, authorized targets, data requirements, and model-hosting choice.

02 / Prepare

Review network paths, permissions, capacity, policies, and required services.

03 / Install and validate

Use the supported deployment process, establish administrator access, and validate the application, inference path, and target connectivity.

04 / Launch and operate

Run a scoped assessment, review its evidence, and establish release, backup, and ongoing operation procedures.

FREQUENTLY ASKED QUESTIONS

Choose placement with the dependencies in view.

Customer infrastructure ownership is not the same as disconnected operation. Confirm the exact configuration your environment requires.

TALK TO ASSAIL

Bring autonomous security testing into your cloud.

Put Ares within reach of private applications and align the deployment with the infrastructure, data, and change controls your team operates.

Your environment. Your objectives. A mission built around both.

https://api.peec.ai/agent-analytics/generic-access-log