v1.8.0 — Developer preview

HandsFree for Chrome 1.8 adds spoken setup, a persistent side panel, more forgiving speech, and recovery that preserves completed work.

• The guided spoken practice shows what was heard and verifies that “open a new tab” created a real tab.
• Natural and relaxed speaking pace, explicit corrections, and short choices for conflicting transcription alternatives.
• Persistent controls with transcript, target, progress, clarification, and Stop.
• Allow a blocked site, choose a missing tab, or edit the command. Eligible resumption runs only unfinished steps after an explicit choice.
• Voice settings first, AI under Advanced, and website-opening macros grouped with routines.
• A lightweight standard package without model weights/runtime; a separate optional local-AI package.

Install or update

Extract handsfree-for-chrome.zip, enable Developer mode at chrome://extensions, and load the folder containing manifest.json. For an update, replace the contents of the existing unpacked folder at the same path and reload. Keep Developer mode enabled. Version 1.8 adds the side-panel permission. See installation instructions.

handsfree-for-chrome-local-ai.zip is the optional model edition. SHA256SUMS lists both archives. The standard package is about 0.5 MiB unpacked, while the local-AI ZIP remains about 212 MiB.

Validation and limits

The application tests, installed-browser command/lifecycle tests, strict TypeScript, lint, build checks, and package checks are described in the validation record. Automated speech tests use synthetic ASR output and fake microphone devices; they do not measure live voice accuracy. OS permission prompts, physical microphone changes, sleep/wake, real battery use, and complex third-party sites need human testing.

AI is off by default. The optional local model failed all 12 exact-plan benchmark cases and remains experimental. Chrome speech recognition may send audio to Google even with local interpretation. Cloud AI uses your own provider key and billing. This is a developer preview, not a Chrome Web Store release.

v1.7.0 — Developer preview

HandsFree for Chrome is an open-source extension for voice control, dictation, and reusable browser routines.

This developer preview includes guided setup and local microphone checks, parameterized routines with import/export and step reordering, and live progress that distinguishes confirmed completions from stopped or skipped work.

Install

Download and extract handsfree-for-chrome.zip. Open chrome://extensions, enable Developer mode, select Load unpacked, and choose the extracted folder containing manifest.json. Do not select the GitHub source ZIP unless you intend to build it first. Existing users should replace their unpacked build and reload the extension.

Validation and limits

The application tests, strict TypeScript, lint, production build, MV3 package checks, and ZIP integrity checks pass. The UI has been exercised with explicitly simulated Chrome responses. Live installed-Chrome microphone behavior, native permissions, and complex third-party websites still require beta testing.

AI is off by default. The bundled small model is experimental and has a documented intent-quality failure. Chrome’s speech service may send audio to Google even when command interpretation is local. Optional cloud AI uses your own key and provider billing.

The ZIP is approximately 212 MiB because it includes local model weights and runtime files. SHA256SUMS contains its checksum. See the README and validation record for details.