Pristan.chat v1 is here

AI agents are already entering companies through SaaS products, internal applications, automation platforms, and experiments built by individual teams. Most of those agents are created and operated outside the systems that employees use to find them.

That creates two practical problems. Companies need a reliable record of which agents exist, who owns them, and who should have access. They also need controls around the information employees send when they use AI at work.

Pristan.chat v1 is ready to address both.

Pristan gives companies an enterprise AI agent catalog and a shared chat workspace. It brings approved agents, access workflows, and interaction controls into one employee experience while allowing the agents themselves to remain on the platforms where they were created.

Catalog agents without pretending to own their runtime

Pristan records agents created in external systems. Each catalog entry can carry an owner, description, business purpose, visibility, access requirements, provider connection, and lifecycle context.

Employees get one place to discover the AI capabilities approved for their role. Administrators get a clearer inventory and a reviewable path for access requests, approvals, revocation, and group-based entitlements.

Apply controls where users interact with AI

When employees use an agent through Pristan chat, the interaction passes through a path the company can configure and review. Depending on the enabled policies and integrations, Pristan can:

• check whether the user is allowed to access the agent;
• inspect prompts and supported files using DLP policies;
• block, warn, redact, or record a policy event;
• retain an audit trail of access and policy decisions;
• attribute usage and cost to users, teams, providers, models, and agents; and
• surface operational, security, and compliance evidence for review.

Those controls apply to the interactions handled through Pristan. If a user accesses the same external agent directly, outside Pristan, that activity is outside Pristan's enforcement path unless a separate integration brings it in.

One approved front door for everyday AI work

Governance works best when the approved path is also the useful path. Pristan combines the catalog with chat, projects, files, knowledge, and supported tools so employees can do real work without jumping between disconnected interfaces.

Companies can connect supported AI providers and change the models behind an approved experience without requiring every user to learn a new front end. Under the BYOK model, provider accounts, credentials, contracts, quotas, and billing remain under the customer's control.

Use Pristan across web, desktop, and mobile

Pristan v1 includes a full Web UI and a responsive Mobile Web UI, giving employees access from a desktop browser or a phone without installing an app.

Native clients for macOS, Windows, iOS, and Android are available in beta. They extend the Pristan experience to the devices employees already use while the web clients remain the primary generally available access paths.

The native clients are built for their respective platforms: SwiftUI on macOS and iOS, WinUI 3 on Windows, and Kotlin with Jetpack Compose on Android. Beta availability means their scope and release cadence may differ by platform as testing continues.

What is included in Pristan v1

• AI agent catalog: agent records, ownership, descriptions, tags, visibility, risk, and lifecycle context.
• Access workflows: requests, approvals, revocation, roles, groups, and entitlements.
• Chat and projects: a consistent place to work with approved agents, files, knowledge, and supported tools.
• Client choice: Web UI and Mobile Web UI, plus beta native clients for macOS, Windows, iOS, and Android.
• Data protection: configurable DLP policies and actions for supported Pristan-mediated prompts, responses, files, and events.
• Audit and compliance support: searchable events, evidence views, retention controls, and export paths.
• Usage and cost visibility: recorded activity by user, team, agent, provider, and model.
• Provider flexibility: customer-managed provider connections and API keys where supported.
• Deployment choice: Docker-based self-hosting and customer-managed deployment patterns.

Built for companies moving from experiments to operations

Pristan v1 is for organizations that no longer want AI adoption to live in scattered links, individual subscriptions, and undocumented team experiments.

It gives IT, security, finance, and business owners a shared system for answering concrete questions: Which agents are approved? Who owns them? Who can use them? What controls apply when people work through Pristan? What activity and cost can we review?

The objective is practical: make approved AI easier to find and use, while giving the company a defensible access and interaction layer around it.