Google launched two new speech models for developers: Gemini 3.8 Live handles real-time voice conversations with reasoning and background tool execution, while Gemini 3.5 Transcribe converts speech to text across 85+ languages with a 4.0% error rate.

• Gemini 3.8 Live can execute API calls in the background while streaming audio responses, handle visual context, and support 97+ languages with accent consistency
• Gemini 3.5 Transcribe achieves 4.0% word error rate (streaming) and 2.6% (non-streaming), with automatic code-switching and custom vocabulary biasing for domain-specific terms
• Extended Thinking variant adds multi-step reasoning capabilities, ranking #1 on Artificial Analysis' Speech-to-Speech leaderboard
• Google's full audio suite includes speech translation (70+ languages), text-to-speech, and music generation all available in the Gemini API

FreeFlow is a free, open-source dictation app for Mac that transcribes speech and cleans up the output using AI, then pastes it directly into any text field. It works with Groq's API by default but lets you plug in any compatible transcription or LLM service.

• Transcription completes in under 1 second using Groq, with optional local model support despite higher latency
• Cleans up filler words and uses app context to spell names and technical terms correctly, with customizable vocabulary lists
• Fully open source under MIT license with no server component—all API calls stay between your Mac and your chosen provider
• Supports both hold-to-talk (Fn key) and tap-to-toggle (Command-Fn) modes, plus custom text-paste commands

The article argues that speaking long, unfiltered “rambles” to an LLM captures your full train of thought and design trade-offs in a way short typed prompts can’t. It offers tips like embracing uncertainties, recording group debates, and voicing emotions to give the model enough context to make decisions that match your priorities. Finally, it shows how to set up a dictation app or transcription pipeline to feed those recordings into your coding agent.

• Rambling long, unfiltered voice memos to Codex (transcribed via Whisper) beats short typed prompts because it captures hidden preferences and context, sharply reducing follow-up questions
• Recording hours-long debates between cofounders gives the AI "ground truth" on team biases/priorities, effectively making it a third team member
• Deliberately voicing uncertainties, false starts, and emotional reactions (not just polished conclusions) reveals constraints that let the AI make on-target decisions

After six months and over 44,000 dictated words with Wispr Flow at 161 wpm, the author tried FluidVoice. It’s an open-source, local Mac app that corrects in real time without an API key and handles slang better. They canceled their paid plan in favor of FluidVoice.

• After 44,414 dictated words at 161 wpm (top 0.1% of Wispr users), the author switched from Wispr Flow to FluidVoice
• FluidVoice is open-source, runs entirely locally on Mac with no API key or cloud dependency, and corrects mistakes in real time
• It handled slang better than expected and matched or beat Wispr Flow's performance, prompting cancellation of the paid Wispr subscription