voicely for mac
Star on GitHub Download
voice context / for your AI

Voice to text for your Mac and AI agent.

Free open-source offline dictation, call transcription with diarization, and local audio-file transcription for macOS — ready for Claude Code, Codex, Cursor, Hermes, and any AI agent on your Mac.

free, forever offline voice-to-text on-device macOS 14+
Mac dictation app Speech-to-text WhisperKit local transcription
Claude Code
> let's pull the auth middleware apart — keep the edge cache, split the session check into its own span, and add a telemetry hook before we touch the rate limiter
Thinking
⌥ + space
the demo
Claude Code
try “fix lint errors”
Your AI knows what you know.
three places

Context comes
from more than
your typing.

Voicely widens the channel between what you know and what your agent can read. Offline dictation, call transcription, and audio-file transcription — each becomes clean text at your cursor or in your agent context.

01 / voice

Offline dictation becomes the prompt.

Press ⌥ + space. Speak freely, with “ums” and restarts. Press again when you're done. Voicely waits for you to finish, transcribes locally, and drops clean voice-to-text at your cursor. No streaming to distract you mid-thought.

02 / calls
recording / team standup 42:18
00:21> let's ship the auth refactor this sprint
00:34> and keep the telemetry hook backwards-compat
00:47> so when we split the middleware…

Your meeting becomes context.

Record any call or room audio on your Mac. Voicely transcribes it cleanly with speaker diarization when available — then you paste the whole meeting into Claude as one block of context, without re-explaining what happened.

03 / files

Your audio file becomes readable.

Drop any audio or video file. Voicely transcribes it locally with WhisperKit — your 90-minute interview is now 12 kB of clean text, ready to hand to your agent. No upload. Nothing leaves your Mac.

where it lives

Wherever your
cursor already is.

Voicely lives one shortcut away — in every app where you type on your Mac. Use it as a dictation app, a speech-to-text layer, or a voice input bridge for AI agents.

claude code
codex
cursor
vs code
openclaw
hermes
slack
telegram
obsidian
notion
how it works / honestly

Transport,
not translation.

Some things Voicely moves directly. Some it packages so you can move them yourself. Either way, nothing is lost between your world and your agent.

mode one

Direct transport.

You speak. Voicely hears you, cleans it up, and lands the text exactly where your cursor is. No middle step, no copy-paste, no window in between.

voice
cursor
mode two

Careful packaging.

You record a call or drop an audio file. Voicely turns it into clean local transcription in one tap. Then you decide where it goes — paste it into Claude, keep it as a note, archive it.

file
text
mode three

Agent handoff.

For longer work, Voicely saves transcripts as files and exposes them through the CLI/MCP path. Claude Code, Codex, Cursor, or Hermes can read the spoken context without another copy-paste loop.

transcript
cli
agent

Nothing leaves your Mac.

offline voice-to-text / on-device / no cloud / no telemetry
faq / search answers

Mac voice-to-text,
without the cloud.

Short answers for the questions people ask before installing a dictation or transcription app on macOS.

Is Voicely free?

Yes. Voicely is free and open source under the MIT license.

Does Voicely work offline?

Yes. Voicely is built for on-device transcription on macOS: dictation, call transcription, and audio-file transcription run locally on your Mac.

Does Voicely upload my audio?

No. The product is designed so audio and transcripts stay on your Mac. There is no cloud transcription service and no telemetry pipeline.

What macOS version is supported?

Voicely targets macOS 14+ and Apple Silicon Macs for the best local WhisperKit transcription experience.

Can I use it with Claude Code, Codex, or Cursor?

Yes. Voicely can insert dictated text at the cursor, and the CLI/MCP flow lets AI agents read saved transcripts as context.

How is it different from Superwhisper, Wispr Flow, and MacWhisper?

Voicely focuses on free open-source macOS voice-to-text, on-device transcription, calls/files, and AI-agent context through CLI/MCP. Check each product's official site for current pricing and platform details.

guides / use cases

Find the workflow
you came for.

Voicely is one app, but people search by job: dictation, offline transcription, call notes, or voice input for AI agents.

mac dictationMac dictation appFree open-source dictation for macOS with text inserted where your cursor already is. privacyOffline dictation for MacOn-device voice-to-text for people who do not want their audio uploaded to a cloud API. voice to textVoice to text on macOSTurn speech into clean text for notes, chats, tickets, code review, and prompts. speech to textSpeech to text on MacLocal speech recognition for notes, prompts, tickets, and long-form context. freeFree Mac dictationA free MIT-licensed route for Mac users who do not want another subscription. open sourceOpen-source dictationInspectable code, local transcription, and a GitHub-first workflow. privatePrivate dictationKeep raw audio on your Mac instead of routing every thought through a hosted API. filesTranscribe audio files locallyMake interviews, lectures, videos, and voice notes readable on your Mac. videoTranscribe video on MacTurn demos, recordings, and videos into text without an upload-first workflow. notesVoice notes to textConvert quick voice notes and longer memos into searchable text. callsCall transcription on MacRecord room/call audio and turn meetings into structured context. meetingsMeeting transcriptionTurn meetings into speaker-aware notes and reusable project context. actionsAI meeting notesMove from transcript to summary, action items, issues, and agent context. whisperLocal Whisper transcriptionA local WhisperKit-oriented path for Mac dictation, calls, and files. whisperkitWhisperKit appMac-native local transcription built around the WhisperKit stack. agentsVoice input for AI agentsUse Voicely with Claude Code, Codex, Cursor, Hermes, and MCP workflows. codingVoice dictation for codingSpeak implementation intent, bug reports, and review notes into coding agents. hermesHermes voice inputUse saved Voicely transcripts as spoken context for Hermes Agent workflows. compareSuperwhisper alternativeA factual comparison for open-source, local, agent-friendly voice-to-text. compareWispr Flow alternativeFor people comparing fast dictation apps with local/open-source workflows. compareMacWhisper alternativeFor file transcription users who also want dictation, calls, and AI-agent context.

Free, forever.

If it's useful to you, say thanks:

or, from the terminal:
$ curl -sL voicely.art/install.sh | sh copy