# Changelog

_Product updates and improvements from Swept AI._

## Playgrounds on Demand: isolated environments in seconds (New)

_2026-06-19T00:00:00Z_

Spin up a fully-isolated, short-lived environment for demos and experiments without touching production.

## Full visibility into every AI application release (Improved)

_2026-06-18T00:00:00Z_

A Changelog tab and structured change summaries give reviewers everything they need before approving a deployment.

## Attach documents and files directly in chat (New)

_2026-06-17T00:00:00Z_

Non-image file attachments now reach the model as-is, without conversion or dropping.

## Per-user usage gauge and dark mode in the chat workspace (Improved)

_2026-06-12T00:00:00Z_

A token budget gauge now persists across chat sessions, alongside a new neutral dark mode and integrated navigation drawer.

## Reports and navigation now carry your brand (New)

_2026-06-11T00:00:00Z_

Platform PDFs and the top navigation automatically display your company logo alongside Swept's.

## Risk tier breakdown gets a clearer chart (Improved)

_2026-06-09T00:00:00Z_

The Risk Tier Distribution panel now uses a percentage bar chart with informational popovers on each tier.

## Budget limits and spend now sync automatically (Improved)

_2026-06-06T12:00:00Z_

A scheduled relay keeps the model gateway and the workbench usage tab in lockstep.

## Redesigned navigation across the platform (Improved)

_2026-06-06T12:00:00Z_

A new top bar and slide-out drawer replace the left sidebar in the workbench and chat.

## Cloud controls dashboard shows live audit events (Security & Compliance)

_2026-06-04T12:00:00Z_

Dashboards now reflect real project data, with a new audit-log feed for continuous visibility.

## Bulk import and document uploads for governance records (New)

_2026-06-04T12:00:00Z_

Populate vendors, AI systems, and data sources via CSV, then attach supporting documents to each record.

## Faster, steadier chat at scale (Performance)

_2026-06-01T10:00:00Z_

Chat sessions stay live across instances, with autoscaling tuned to your traffic.

## Locked-down controls for the chat workspace (Security & Compliance)

_2026-06-01T09:00:00Z_

Configuration is fixed at startup, with higher-risk features off by default.

## A Swept-branded chat workspace (New)

_2026-05-31T09:00:00Z_

A managed, on-brand chat experience built into the platform.

## One gateway to leading AI models (Integration)

_2026-05-30T11:00:00Z_

A single endpoint connects your apps to Google Gemini and Anthropic Claude.

## Web search and reasoning in chat (New)

_2026-05-30T10:00:00Z_

Models can search the web and reason through harder questions before answering.

## Examination-ready compliance packages (Security & Compliance)

_2026-05-30T09:00:00Z_

Generate audit-ready documentation packages on demand.

## Token spend and budget controls (New)

_2026-05-26T09:00:00Z_

Set budgets by user or group, with in-app alerts and forecasting.

## Stream supervision events to your stack (New)

_2026-05-24T09:00:00Z_

Send live supervision events to your own systems through signed callbacks.

## Enterprise SSO and per-group access (Security & Compliance)

_2026-05-21T09:00:00Z_

Sign in with your identity provider, with access controlled per group.

## Ongoing platform security hardening (Security & Compliance)

_2026-05-18T09:00:00Z_

Proactive patching keeps the platform current against known vulnerabilities.

## An AI governance suite (New)

_2026-05-15T10:00:00Z_

Risk tiers, approval workflows, vendor management, and drift monitoring in one place.

## One unified workspace (Improved)

_2026-05-15T09:00:00Z_

Shared navigation and theming across Workbench, Cloud Lock, and chat.

## Deploy entirely in your own cloud (Security & Compliance)

_2026-05-06T09:00:00Z_

Run Swept inside your own private cloud with full network isolation.

## Cloud Lock: one view of your AI infrastructure (New)

_2026-04-23T09:00:00Z_

A single pane for infrastructure health, security posture, and inventory.

## Synthetic Knowledge Base (New)

_2026-04-21T09:00:00Z_

Build grounded knowledge from the systems you already use, with built-in QA.

## Stronger evaluation with fairness checks (Improved)

_2026-04-13T09:00:00Z_

Multi-source evaluation, independent verification, and fairness checks.

## 100+ new AI agent and help-desk integrations (Integration)

_2026-04-02T09:00:00Z_

Connect agents across 100+ platforms through a new adapter catalog.

## Performance tuned for large deployments (Performance)

_2026-04-01T09:00:00Z_

Query optimizations and bulk operations for high-volume environments.