Short answer: DaVinci Resolve has no built-in dictation, so use a system-wide Mac dictation app that types at your cursor. Click into any Resolve text field, hold your hotkey, speak, release. Marker notes and Media Pool metadata are where it pays off most, and you must mute playback first.
DaVinci Resolve is an editor, a colour suite, an audio workstation, and a delivery tool. What it is not is a writing app, which is exactly why most editors write so little inside it. Markers get named with a single word. Clip descriptions stay empty. Comments on the timeline never happen. Not because the information is worthless, but because typing it means letting go of the keyboard shortcuts you have spent years learning, moving the pointer to a small text field, and typing with the hand that was on J K L a second ago.
That is a friction problem, not a discipline problem, and it is the kind of friction voice removes cleanly. This guide covers how to set dictation up in Resolve on a Mac without breaking your shortcut muscle memory, which surfaces are actually worth speaking into, and the one mistake that will make it seem broken on your first attempt.
There is no dictation inside Resolve, and that is fine
Blackmagic ships transcription features in Resolve, but they are aimed at a different job: turning the audio in your timeline into text, for subtitles or for text-based editing of interviews. That is transcription of your footage. It does not help you write a note about the footage.
What you want for writing is system-wide dictation: a tool that types wherever your cursor is, in any application, without needing a plugin or an integration. That is how Voice Keyboard Pro works on the Mac. It lives in the menu bar, you hold a hotkey, you speak, you release, and the words appear at your text cursor. As far as Resolve is concerned, something typed. There is nothing to install into Resolve, no scripting, and no difference between the free version and Studio, because the dictation never touches Resolve's internals.
Setup: pick a hotkey Resolve has not already claimed
This is the whole setup, and it takes about a minute, but it deserves care in Resolve specifically.
Resolve, like most professional NLEs, uses bare single-key shortcuts aggressively. J, K, and L for shuttle. I and O for in and out. B for blade. A for the selection tool. M for markers. These are why the app is fast, and they are why you should not choose a bare letter or a simple modifier combination for your dictation hotkey. Pick something Resolve will never send anywhere useful.
- Good choices: a held modifier such as Right Command or Right Option, a function key, or a modifier plus a key that Resolve leaves alone.
- Bad choices: anything single-letter, anything in the transport cluster, and anything you use in the colour page while grading with one hand on the panel.
- Verify before you commit: open Resolve, put focus on the timeline, and press your candidate hotkey a few times. If nothing in the timeline moves, changes tool, or drops a marker, you are clear.
The second half of setup is a habit rather than a setting: the text must have somewhere to land. Dictation types at the cursor, so if the cursor is not in a text field, the characters go to whatever has focus, and in Resolve that can mean triggering shortcuts. Click into the field first, then speak. After a day this becomes automatic in the same way that clicking into a rename field is.
The program audio trap
Here is the mistake that makes people conclude dictation does not work in an editing app. You are reviewing a clip, playback is running through your speakers, you hold the hotkey and speak, and the result is a mess of your words tangled with dialogue from the timeline.
Your microphone cannot tell the difference between you and your programme audio. Neither can any transcription engine. The fix is trivial once you know it: stop playback before you speak, or monitor through headphones. Editors who work on headphones never encounter this problem at all. Editors on speakers hit it in the first five minutes and often blame the app.
Where dictation actually pays off in Resolve
Not every text field in Resolve is worth speaking into. Some are worth a great deal.
1. Marker notes, which are the whole ballgame
Markers are the highest-leverage surface in the application for voice, by a wide margin. Drop a marker on the timeline or on a clip, open it, and you get a name field and a notes field. Almost everyone fills in the name, badly, with something like "fix" or "audio". Almost nobody fills in the notes, because the note you actually want to leave is a sentence and typing a sentence costs more than the note is worth in the moment.
Except the note is worth a lot. "Fix" tells you nothing three days later. "Camera B goes soft on the push in from about the two second mark, either use the wide or ask if there is a second take" tells you everything, and it takes six seconds to say. This is the single change that makes the biggest difference to a Resolve project, and it is available the moment you have a hotkey.
The workflow that works:
- Play until you see the problem, stop.
- Drop the marker.
- Open the marker dialog and click into the notes field.
- Hold, speak one or two sentences, release.
- Close and keep going.
You will not go back to one-word markers.
2. Media Pool metadata during the logging pass
Logging is the most tedious part of any project with real footage volume, and it is almost entirely a writing task pretending to be an editing task. Description, comments, scene and shot notes, keywords: all of it is prose, all of it is typed grudgingly, and all of it determines whether Smart Bins and search are useful later.
Dictating the description field while you watch each clip changes the economics of logging. You watch, you speak what you saw, you move on. The pass takes less time and produces richer metadata, which is the rare combination where the faster method also yields the better result. If you are doing a long logging session, treat it as a batch: get all the descriptions in one sitting rather than scattering them through the edit.
3. Subtitles and caption corrections
If you are working with a subtitle track, most of what you do is correcting text rather than writing it from scratch. Speaking a full corrected line is often faster than surgically fixing three words in the middle of it, especially for the long lines. This is one of the few places where you should consider replacing the whole line by voice rather than editing in place.
4. Text+ and titles
Lower thirds, credits, chapter cards, and social captions all live in text fields, in the Inspector or in Fusion. Longer title copy, especially the descriptive kind that runs to a sentence or two, is worth speaking. Short names and single words are usually faster typed. The dividing line is roughly: if it is a phrase you would say out loud, speak it; if it is a proper noun and nothing else, type it.
5. Everything around the edit
This is the part editors underestimate. Most of the words in an editing job are not inside the editing app at all. They are in the client email, the round of notes you reply to, the YouTube description, the scope clarification, the invoice line item, the message explaining why the deadline moved.
A system-wide dictation tool covers all of it with the same hotkey, which is the argument for using one rather than an app-specific feature. When you finish the cut and switch to the browser to write the video description, nothing changes about how you work. We covered the publishing side of that in dictating YouTube descriptions and comments, and the broader profession view in voice typing for video editors.
Where not to use it
Being honest about this is what keeps the tool useful.
- Timecodes. Type them. Speaking digit strings with colons is slower and error-prone, and the field usually wants a specific format.
- Node names and clip names with codes in them. Anything with model numbers, underscores, or camera and card identifiers is a keyboard job.
- Fusion expressions and any scripting. Syntax is not prose. Do not fight this one.
- File paths and delivery filenames. Same reason. A mistake here is expensive and silent.
- Fine parameter tweaks. Obviously, but worth saying: dictation is for text, not for the colour page.
The pattern across all five is the same. Dictation is excellent at language and mediocre at structured identifiers. Resolve happens to contain a lot of both, so the discipline is knowing which field you are in.
Making it accurate on your kind of project
Two things reliably improve results for editors, and neither is about speaking more clearly.
The first is vocabulary. Post-production runs on terms that are not ordinary English: codec and format names, camera bodies, LUT names, client and brand names, agency shorthand, the internal name of the project. Those are the words a transcription engine has the least reason to guess correctly, and no amount of enunciation fixes it, because the problem is not your diction. It is that the word is rare.
Voice Keyboard Pro's Smart Vocabulary is a personal dictionary with replacement rules, so you add the terms once and they stop being a problem across every project. Client names and product names are the highest-value entries; add them at the start of a job and the rest of the job gets quieter.
The second is accepting that dictation produces a draft. For marker notes and metadata, the draft is the finished product and you should not spend a second polishing it. For anything a client will read, run a quick cleanup pass. The two-pass habit is the difference between people who stick with dictation and people who quit in week one, and it is laid out fully in our guide to editing dictated text. Voice Keyboard Pro also has Voice Edit, where you speak the change you want rather than retyping the sentence, which suits the case where you have dictated a long client reply and want to soften the tone of one paragraph.
A realistic session
Here is what a Resolve day looks like with dictation folded in, without reorganising how you edit.
Ingest and logging. Headphones on. Watch each select, click the description field, speak two sentences about what is in the shot and whether it is usable. Twenty clips takes a fraction of the time it would typed, and the bins are searchable afterwards.
Assembly. Hands on the keyboard, shortcuts as normal, no voice at all. Editing is a keyboard and mouse task and should stay one.
Review pass. This is the heavy voice section. Play, stop, marker, speak the note, continue. You end up with a timeline annotated in full sentences, which is what makes the next session start fast instead of starting with ten minutes of re-watching to remember what you meant.
Client round. Notes come back, usually as a list. Replies to notes are prose and belong to your voice. So does the email confirming what you changed.
Delivery and publishing. Render settings typed. Description, chapter list, and social copy spoken, in the browser, with the same hotkey.
The total speaking time is small and it lands on the exact tasks that normally get skipped or done badly.
Frequently asked questions
Does this work in the free version of DaVinci Resolve?
Yes, identically. Dictation types into text fields at the operating system level, so the free version and Studio behave the same way. There is no integration to license.
Is this the same as Resolve's transcription feature?
No, and the two solve different problems. Resolve's transcription turns the audio in your timeline into text so you can subtitle or edit from a transcript. Dictation turns your voice into text wherever your cursor is, which is what you want for notes, metadata, titles, and email. Many editors use both.
Will my dictation hotkey break my Resolve shortcuts?
Only if you choose one Resolve already uses. Pick a held modifier such as Right Command, or a function key, and test it in the timeline before you commit. Resolve leans heavily on bare single letters, so avoid those entirely.
Why is my dictation picking up the video's audio?
Because your speakers are feeding your microphone. Stop playback before speaking, or monitor on headphones. This is the most common first-day complaint in any editing application and it has nothing to do with the dictation tool.
Can I dictate in the Fusion or Fairlight pages?
You can dictate into any text field on any page, since it types at the cursor. In practice Fusion text tools and Fairlight marker notes are the useful targets. Expressions, node names, and channel labels are better typed.
Does it work offline?
No. The transcription engine needs a connection. If you edit in a location without network access, keep macOS's built-in dictation available as a fallback for that situation.
What happens to what I say?
Voice Keyboard Pro's servers store only operational pings. No audio and no transcript content. The text goes to your cursor and stays in your project. If you work under an NDA, the usual judgement still applies about what you say out loud in a shared room, which is a separate question from where the data goes.
Is it worth it if I only cut short videos?
The marker-notes benefit scales with project length, so a two minute social edit gains less. The surrounding writing does not scale down, though. If you send client emails and write descriptions, that part is the same whether the cut is two minutes or forty.
The bottom line
Resolve is one of the most capable applications on the Mac and one of the least pleasant places to type, which is why so many projects carry markers named "fix" and clip descriptions that are empty. That gap is not a sign that the information does not matter. It is a sign that the cost of recording it is higher than the moment allows.
Speech runs at 130 to 150 words per minute against roughly 40 for average typing, and the notes you skip are exactly the ones in that gap. Keep your hands on the shortcuts for the edit. Give the notes, the metadata, and the client email to your voice. If you also cut in Final Cut, the same hotkey works there, and we wrote that one up separately in dictating in Final Cut Pro.
Voice Keyboard Pro has a free tier with daily limits, and Pro is $4.99 a month or $34.99 a year. Try it on the next marker you were about to name "fix".