v0.2.85 v1.0.53

high risk

mlc-ai/web-llmSep 8, 2026 → Sep 9, 20262 days between releases

What's New

Major version jump introduces workflow automation and passkey features alongside Copilot and browser tooling fixes.

New Features

  • Gate passkey 2FA behind enterprise plan
  • User-selectable browser type for workflows
  • Workflow retry policy settings
  • Attempt-aware run views
  • Screenshot capture failure observability

Improvements

  • Copilot build tests run in matching browser
  • Screenshot capture refinement
  • CPU request fairness adjustments
  • Copilot drafting block naming
  • File upload result classification

Bug Fixes

  • Raw-CDP screenshot rescue on persistent-context browsers
  • Type forced click focus fallback
  • Authenticator key editability
  • Fill tool behavior in extraction blocks
  • Visibility check reporting in Task V3

Upgrade Notes

The v1.0.53 notes do not reference the v0.2.85 runtime updates. Review workflow configurations for new passkey gating and retry policies. Ensure browser automation settings account for the new user-selectable browser type.

The Two Releases

v0.2.85

v0.2.85

Sep 8, 2026· 1 month ago

This release updates the bundled TVM runtime and introduces synchronous OPFS access handles. It improves stability by preventing hangs and rejecting invalid parameters, while ensuring OpenAI-compatible API timestamps and preserving worker model state.

Full release notes
v1.0.53

Release v1.0.53

Sep 9, 2026· 1 month ago

A substantial update introducing new features like passkey 2FA gating, user-selectable browsers, and workflow retry policies, alongside numerous bug fixes for the Copilot, Task V3 engine, and browser automation tools.

Full release notes