TypeWhisper/typewhisper-mac v1.5.0-rc1

on GitHub

TypeWhisper 1.5.0-rc1 opens the 1.5 release-candidate line with app-aware dictation insertion, virtual audio input support, broader number normalization, new speech providers, and a focused reliability pass across plugins, model handling, workflows, and release metadata.

Highlights

• Added app-aware smart insertion so dictation output better respects the surrounding text, sentence position, trailing spaces, terminal paste behavior, and target-app context.
• Added virtual audio input device support and hardened media pause/resume handling, recording start feedback, mouse-button shortcuts, and Safari/fullscreen indicator behavior.
• Expanded number normalization across generic transcription output, multilingual number words, Dutch number words, English ordinals, and English digit sequences.
• Added new bundled speech providers and plugin releases, including Gemini speech transcription, Cartesia speech transcription, Sber SaluteSpeech, OpenRouter speech-to-text, Reson8, OpenAI-compatible profiles, and the experimental SenseVoice line before its removal from the bundled set.
• Added Japanese localization and dictation support, ordered language hints, recent transcriptions in the workflow palette, source progress for file transcription, recorder-specific engine overrides, and a short-dictation workflow AI-skip setting.
• Added commercial Cloud Folder Sync, the standalone setup wizard, README screenshot automation, plugin update-state diagnostics, repair actions, and clearer issue-template diagnostics guidance.

Reliability and Fixes

• Fixed smart insertion mid-sentence edge cases, terminal clipboard restore races, trailing-space insertion, middle-mouse pass-through, indicator visibility over fullscreen apps, and overlay live transcript preview expansion.
• Improved long file transcription progress, Groq uploads, Apple Speech first-use model selection, selected Apple Speech model restoration, custom workflow fine-tuning prompts, and workflow prompt casing diagnostics.
• Hardened cloud LLM behavior by honoring plugin model selection, parsing Gemini array errors, and avoiding legacy global model poisoning.
• Updated Soniox async transcription to model v5 and disabled dictionary-term prompts for Qwen3 ASR where they do not apply.
• Treated Reson8 as a single-language engine, added configurable LLM request timeout support to the OpenAI Compatible plugin, and required TypeWhisper 1.5 for the Gemini plugin.
• Fixed Cartesia plugin settings and prevented plugin releases from taking over the repository-level GitHub "latest" marker.

Testing Focus

• Verify app-aware insertion in plain text, mid-sentence, terminal, and rich-text targets.
• Verify virtual audio input recording, media pause/resume, mouse-button shortcuts, and indicator behavior in Safari fullscreen and other fullscreen apps.
• Verify number normalization for multilingual number words, Dutch, English ordinals, English digit sequences, and API/CLI dictionary corrections.
• Smoke-test Gemini, Cartesia, Sber SaluteSpeech, OpenRouter STT, Reson8, and OpenAI-compatible profile setup paths.
• Verify file transcription progress, recent-transcription workflow palette behavior, ordered language hints, Japanese dictation, and short-dictation AI-skip workflows.
• Confirm v1.5.0-rc1 publishes on the release-candidate Sparkle channel, remains a GitHub prerelease, and does not update Homebrew or the stable website release path.