Skip to content

XIPU AI Bridge v0.3.0

Choose a tag to compare

@github-actions github-actions released this 09 Oct 06:36
· 7 commits to main since this release
v0.3.0
fac5174

What's new

  • Add a first-run setup guide that creates and verifies a dedicated XIPU AI conversation without generating a response.
  • Recover interrupted setup with read-only checks, cancellation, and manual configuration through Settings. Uncertain creation attempts are not automatically repeated.
  • Unify branding across the popup, setup guide, and settings.
  • Accept clean school stream completion without requiring an optional [DONE] marker, while retaining errors for incomplete frames and failed reads.
  • Return HTTP 400 with unsupported_image_model when a request contains images but the selected model does not advertise image support.
  • Add Text-only image history, disabled by default. When enabled, earlier answered images are replaced with an unavailable marker while existing text descriptions are retained. New images still require image support.

Install or update

Download the native archive for your operating system and CPU, plus xipu-bridge-extension.zip. Extract both and follow the included INSTALL.md.

Existing installations should rerun the native installer, replace the extension files, reload the extension in Chrome, and refresh the signed-in XIPU AI tab. Existing native settings and API credentials are preserved. Enable Settings → General → Text-only image history only if you want that behavior.

Verification and limits

Automated checks run on macOS, Linux, and Windows, including synthetic browser/native integration and package validation. Live service behavior and browser integration on every platform require separate verification. Native binaries are unsigned and not notarized. Omitted images cannot be inspected by the model; only retained descriptions remain available.

Full changelog: v0.2.0...v0.3.0