Skip to content

AI & automation features

Capabilities, clearly labelled.

Every feature shows its real status: Available, Beta, In Development, Planned or Research.
IntelligenceBeta

Cloud models

Chat with supported cloud AI models through the web workspace.

IntelligenceIn Development

Local models

Run compatible models on your own device without cloud inference.

IntelligencePlanned

Hardware-aware recommendations

Suggests models suited to your available memory and compute.

IntelligenceBeta

Document understanding

Ask questions about supported attached files.

IntelligencePlanned

Research assistance

Explore sources, extract key points and organise findings.

AutomationIn Development

File organisation

Propose and apply reviewed folder changes on desktop.

AutomationPlanned

Workflows

Repeatable, permission-controlled multi-step tasks.

AutomationPlanned

Approved connectors

Connect authorised apps with clearly scoped access.

AutomationIn Development

Task planning

Turn instructions into a restricted, reviewable action plan.

AutomationPlanned

Reversible operations

Undo supported changes after execution.

SecurityIn Development

Independent policy enforcement

A permission engine that sits outside the model's authority.

SecurityIn Development

Permission management

Scoped, expiring permissions you can inspect and revoke.

SecurityPlanned

Audit history

A readable record of decisions and executed actions.

SecurityResearch

Prompt-injection defence

Layered handling of untrusted content to reduce injection risk.

SecurityResearch

Risk scoring

Approvals that adapt to the sensitivity of an action.

PersonalisationPlanned

AI memory

Useful context with privacy and provenance controls.

PersonalisationPlanned

Workspaces

Separate contexts for different projects and roles.

PersonalisationBeta

Model preferences

Choose the model used for each conversation.

PersonalisationPlanned

Local/cloud policies

Rules for when data may leave your device.