Three differences, in one sentence each.
Local Whisper. System-wide capture. Auto-enriched notes. That's it.
The Whisper plugin and ThoughtMic both put transcribed voice into your Obsidian vault. The plugin does it brilliantly, inside the editor, free, using OpenAI's Whisper API. ThoughtMic differs on three specific axes — and if none of them matter to you, the plugin is the right choice.
1. Local Whisper, no API key
The plugin uses OpenAI's Whisper API: you bring your own key, your audio uploads to OpenAI for transcription, and you pay per minute (a few dollars a month for typical use — not a huge issue for most people). ThoughtMic ships Whisper Large V3 Turbo and runs it locally on Apple Silicon with Metal acceleration. No API key. No upload. No per-minute cost. If you'd rather not have audio leave your machine — for privacy, for principle, or just to skip key management — that's the difference.
2. System-wide capture, not just inside Obsidian
A plugin, by definition, runs inside its host app. The Whisper plugin works while you're in Obsidian — that's the entire point of being a plugin, and it's what makes installation so frictionless. ThoughtMic is a system-wide Mac app: press the hotkey in Slack, Mail, your IDE, a browser, your terminal — text lands at the cursor wherever you are. And in parallel, a structured note can land in your vault from the same dictation. If "I want voice everywhere on my Mac, with notes also flowing to my vault" describes you, that's a different shape of tool.
3. Auto-titled, tagged, backlinked notes — not raw text in the current file
When you dictate with the plugin, the transcribed text lands in your currently open note. That's the right behavior for in-editor dictation, and it's what most people want from a plugin. ThoughtMic treats every dictation as a candidate note in your vault: it generates a title from the content, suggests tags pulled from your existing tag taxonomy, and resolves [[wikilinks]] against notes you already have. If you want voice to feed your knowledge graph automatically rather than dropping into whatever file happens to be open, that's the differentiator.
A note on transcription quality and speed
Whisper is Whisper. The plugin uses OpenAI's hosted version; ThoughtMic uses the same architecture (Whisper Large V3 Turbo) running locally. Quality is essentially identical for clean audio. Latency is comparable on M-series Macs (~700–1500ms end-to-end), with the plugin sometimes faster on a great connection because OpenAI's hosted inference is heavily optimized, and ThoughtMic sometimes faster on a flaky one because there's no network round-trip. Neither is the streaming sub-300ms experience of cloud-streaming dictation tools — and getting there means giving up local-first.
"Nik's Whisper plugin is excellent work and the standard for in-Obsidian voice capture. Keep using it if you're happy. ThoughtMic exists for people who want local transcription without an OpenAI key, voice capture from anywhere on their Mac, and automatic tags and backlinks rather than raw text into the current note. Different tools for different needs."