WAI Docs Wed Aug 19 13:22:37 EDT 2026
List
Quick Start
Quick Start
User Guide
User Guide
Policies - GuardRails
Policies - GuardRails
Witness Anywhere: Remote Device Security
Witness Anywhere: Remote Device Security
Witness Attack
Witness Attack
Administrator Guide
Administrator Guide
404
404
WitnessAI Release: April 23, 2026
Release Date: April 23, 2026
Note: New and updated features may not be immediately available on your dedicated deployment due to participation in limited availability releases, beta programs, or regional configurations.
If you don't see a feature you expect, please reach out to your Customer Success team — we're happy to help.
New Features
Writer.com
Writer.com Observability: Full prompt and response capture is now supported for Writer.com (app.writer.com). Observability, policy enforcement, and data protection controls are now active for Writer.com traffic.
Feature Updates
ChatGPT Thinking Mode: Added support for prompt edits made while ChatGPT is in Thinking mode, so when a user edits a prompt mid-response, the updated request is updated in Conversations.
Performance, Reliability, & API Updates
Provider Response Time: When the LLM timing header is unavailable, the value is populated by a fallback metric, eliminating blank or zero values in the console.
VSCode Anthropic Endpoint: Reduced latency for users working with Claude models in VSCode.
Fixed Issues
Data Protection
- Fixed an issue where data protection warning messages were not displayed in Google Sheets (Gemini) for some prompts.
Agentic Policy
Note: Agentic Policy features are currently available behind a feature flag. If you are part of a Limited Availability Release (LAR) and expect to see this, please contact Customer Success.
- Fixed an issue where bulk-adding MCP Servers or Tools to an Agentic Policy only saved the first item added.
- Fixed an issue where saved Agentic Policies were deleted after changing the Allowlist or Blocklist mode from the dropdown.
Conversations & Alerts
- Fixed an issue where some conversation export requests returned no results.
- Fixed an issue where single VSCode prompts (Claude) were displayed as multiple separate prompts in the Conversations view.
- Fixed an issue where some prompts generated duplicate Alerts.
Console Display
- Fixed an issue where markdown tables in AI responses were displayed as raw text instead of formatted tables on the conversation detail page.
Session Management
- Fixed an issue where session information was not properly validated before requesting a documentation token, which could cause unnecessary authentication calls.
Known Issues
No known issues identified in this release.
Breaking Changes
No breaking changes identified in this release.