Blog
Thoughts on voice, AI, and the craft of making tools that disappear — from a small team of algorithm researchers and daily AI-product users.
Omni-modal voice typing: multimodal understanding, MoE, and streaming text output
Omni-modal voice typing needs a split pipeline, streaming audio, and Mac-local context. A look inside Loqua's voice stack.
Compare
Loqua vs Wispr Flow: a Mac-first Wispr Flow alternative for context, coding, and privacy
An honest Wispr Flow alternative comparison: Loqua vs Wispr Flow on Mac context, coding workflows, privacy controls, latency, and pricing.
Loqua vs Typeless: a Mac-native Typeless alternative for context, coding, and depth
A Typeless alternative comparison for Mac: Loqua vs Typeless on context awareness, coding workflows, privacy posture, and pricing.
How-to
Mac meeting notes voice: from voice to done with notes and action items
Mac meeting notes voice workflow: capture notes, action items, Notion bullets, Things tasks, and follow-ups without leaving your call.
How to dictate code on Mac: a complete guide for Cursor, VS Code, and Claude Code
How to dictate code on Mac with Loqua — camel-case identifiers, commit messages that pass review, PR descriptions in your voice. Six worked examples + setup.
Voice typing for AI coding: voice prompt Cursor and Claude Code without typing
A practical voice prompt Cursor + Claude Code guide: structured prompts, file context, Cursor and Claude Code workflows, and mixed EN+Chinese.
Hands-free dictation for writers: how to draft 3000 words of novel, essay, or long-form in one sitting
A dictation for writers workflow: cleaner drafts, character dictionaries, app-aware formatting, editing by voice, and long-form practice.
Engineering
Reinforcement learning voice typing: GRPO, DPO, and on-policy distillation in our voice stack
Reinforcement learning voice typing helps Loqua improve rare terms, latency, and natural output after supervised training stops paying off.
Multimodal voice recognition: building a listener that sees what you see
Multimodal voice recognition helps Loqua resolve homophones, code identifiers, and app context by combining audio with local screen cues.
Audio event detection dictation: sounds with meaning beyond words
Audio event detection dictation is a research direction at Loqua: laughter, pauses, and sound cues as optional context for structured writing.
Voice typing architecture: inside Loqua's three-model voice typing stack
A blog-level look at Loqua's voice typing architecture — why three purpose-trained models (speech recognition, language intelligence, screen context) beat one.
Voice meets vision: how omni-modal models unlock multimodal voice typing
Why multimodal voice typing needs audio plus screen context, what the omni-modal research shows, and what Loqua applies on Mac.
Private voice dictation Mac edition: how Loqua's hybrid voice typing stack keeps your data on your side
A private voice dictation Mac architecture: what Loqua keeps local, when optional cloud processing can be used, and how privacy boundaries actually work.
Productivity
Voice for thinking with AI: why your keyboard is the wrong tool
Voice for thinking with AI helps preserve half-formed ideas before the keyboard compresses them into brittle prompts.
Voice first workflow: a day in our voice-first workday
Voice first workflow notes from Loqua's founder: inbox, standup, code review, spec drafting, Slack, and journaling.
Voice productivity stack: 9 tools we actually use to write, ship, and think
Voice productivity stack guide: Loqua, Claude Code, Cursor, Obsidian, Granola, Linear, Raycast, Things 3, and Spark.
Voice typing workflows: 8 underrated workflows for daily AI work
Voice typing workflows for AI work: commits, PRs, Linear issues, replies, brainstorms, specs, code comments, and standups.