Media management for documentary editors

Your footage is logged before
you open the edit.

HeyMonkey Library ingests your interviews and b-roll read-only, watches and transcribes every clip, identifies who's speaking, scores the shots, and hands DaVinci Resolve a Media Pool that's already organised, tagged, and searchable β€” down to the spoken word.

🐡 Sign up for beta testing β†’ See how it works
Read-only ingest Β· camera-card safe Your own AI keys Your own database Runs offline
HeyMonkey β€” Riverbank (Documentary)
All 148πŸŽ™οΈ Interviews 22🎬 B-Roll 126
Keyword🧠 Meaning
πŸ’¬ Ask the footage
β˜… Min score38%
148 clips β˜… Picks Β· 31 βœ“ Indexed Β· 148 πŸŽ™οΈ Diarized Β· 22 ⊘ Rejects Β· 9
Interviewβ˜…πŸŽ™οΈβ˜… 92
A012_C004_Naledi.mov
02:14Β· πŸ—£ Naledi M.
B-Rollβ˜…πŸŒ‡β˜… 88
DJI_0421_riverbank.mp4
00:41Β· golden hour
B-Rollβ˜†πŸŒΏβ˜… 74
C087_wetland_wide.mov
01:07Β· calm
Interviewβ˜…πŸ‘€β˜… 90
B004_C002_Thabo.mov
03:52Β· πŸ—£ Thabo K.
01

AI quality score on every clip β€” sharpness, steadiness, exposure β€” so the best takes rise.

02

Pick / reject flags and one-click triage pills that filter the whole bin instantly.

03

People & transcripts attached to each clip β€” search by who's on screen or what they said.

The Library, recreated β€” the real interface, same controls and colours.

The dead hours before the cut

Logging, syncing and transcribing eat the days
before you make a single edit.

On a documentary you drown in footage β€” dozens of cards, multiple cameras, hours of interviews. Before the story can start you're renaming clips, hunting the soundbite, matching timecode, and typing transcripts. HeyMonkey does that pass for you, in the background, and leaves the creative decisions where they belong: with you, in Resolve.

148
clips watched, described & scored β€” unattended
22
interviews transcribed & speaker-labelled
1
click to send it all into Resolve, organised
0
original files touched β€” read-only, always
How it works

Five stages, from camera card to cut.

Each stage runs on its own β€” or leave it going and come back to a fully-logged project.

01
πŸ“₯

Ingest

Point it at your cards. Originals are read in place and never moved or modified β€” camera-card safe.

02
🧠

Index

AI watches each clip: a plain-language description, shot type, mood, faces, and a quality score.

03
πŸŽ™οΈ

Understand

Transcribe every interview, identify each speaker, and surface soundbites, emotions and claims.

04
πŸ—‚οΈ

Organize

Multicam & timecode Scene Assets, duplicate stacks, smart bins, and a coverage gap list.

05
🎬

Hand off

Send to Resolve β€” metadata, bins, markers β€” plus the in-Resolve panel, FCPXML, ALE & Story Bible.

Find anything

Search two ways β€” or just ask.

Keyword is an instant match across descriptions, tags, people and transcripts. 🧠 Meaning finds a moment by sense β€” "quiet resilience", "a turning point" β€” even with no shared words. And Ask the footage answers a plain question, grounded only in this project, citing the exact clip and timecode.

  • βœ“Every answer cites a clip + timecode you can click to play the moment.
  • βœ“Turn an answer straight into selects or push it to Resolve.
  • βœ“Scope a question to a person, a tab, or the whole project.
πŸ’¬ Where does anyone talk about the river changing over time?
Two speakers address it directly. Naledi describes the seasonal floods retreating year on year A012_C004 Β· 01:12, and Thabo ties the change to upstream farming B004_C002 Β· 02:47. A third mention is more oblique, framing it as loss A030_C001 Β· 00:38.
↳ Who blames upstream farming?↳ Any hopeful takes on the river?↳ Show b-roll of flooding
Every clip, at a glance

A card that already knows the shot.

Poster, AI quality score, pick/reject flag and category β€” readable in a scroll. Right-click any clip (or the theatre player) for the full toolset: re-tag, isolate a mic, identify speakers, deep-index, link a multicam angle, or open the contributor dossier.

  • βœ“Score ribbon from a local sharpness/stability/exposure pass β€” no cloud needed.
  • βœ“πŸŽš Audio track isolation β€” transcribe the lav, not the boom, per clip.
  • βœ“Theatre player plays a fast proxy or the original, your call.
Interviewβ˜…πŸŽ™οΈβ˜… 92
A012_C004_Naledi.mov
02:14Β· πŸ—£ Naledi M.Β· hopeful
This clip
πŸŽ™οΈ Set as Interview
🎬 Set as B-Roll
Audio & speakers
🎚 Audio track (isolate mic)
πŸ—£οΈ Identify speakers
πŸ‘€ Contributor dossier
Analyse
🧠 Deep index this clip
πŸŽ₯ Link multicam angle
Who said what

Speakers, identified β€” and balanced.

On-device diarization labels every line by speaker, links names to faces, and rolls up a talk-time balance across the whole project. Jump to any person's every line, or export their selects in a click.

  • βœ“Runs offline on your GPU or CPU β€” no audio leaves the machine.
  • βœ“Names flow into the cast list, soundbites, and the Story Bible.
  • βœ“Speaker lines are searchable and exportable as a selects reel.
TALK-TIME BALANCE Β· RIVERBANK
Naledi M.41%
Thabo K.28%
Grace P.19%
Narrator12%
00:47  Naledi: "When I was a girl the water came right up to here…"
Sync, solved

Scene Assets from timecode.

HeyMonkey reads each clip's embedded start timecode and groups jam-synced A/B-cam and external audio from the same shoot into one Scene Asset β€” frame-exact, no audio guessing. Or audio-sync claps the old way when there's no matching TC.

  • βœ“Drop-frame aware; handles mixed frame rates across cameras.
  • βœ“Groups only within the same shoot day β€” no false midnight matches.
  • βœ“Becomes a multicam group Resolve and the AI editor both understand.
🎬 SCENE 3 · 2026-05-14 · 25p
A Β· MASTERA012_C004.mov10:14:02:00
BB004_C002.mov+0.00s Β· full overlap
CZOOM_0031.wav+0.04s Β· covers 0:00–2:11
βœ“ 3 clips groupedmethod Β· timecodeconfidence Β· high
Assemble

Build the story on the page first.

Highlight transcript passages or soundbites and drag them into an ordered sequence β€” a paper edit. Export it as an FCPXML cut-list that lands the clips on your Resolve timeline in narrative order, or as a radio-cut MP3 to hear the flow.

  • βœ“Reorder by drag; running duration as you build.
  • βœ“Every passage keeps its clip + in/out β€” the cut-list is frame-true.
  • βœ“Best soundbites, emotions and coverage-gaps feed the assembly.
SOURCE Β· SOUNDBITES
β˜… "…the water came right up to here" 0:47
β˜… "…upstream, the farms took it" 2:47
ASSEMBLY Β· 3:12
1 Naledi β€” childhood river
2 Thabo β€” the cause
3 Grace β€” what's at stake
Export FCPXMLRadio cut
Built for DaVinci Resolve

The Library lives where you edit.

HeyMonkey isn't a walled garden β€” it exists to make your Resolve project better. Three ways in, from a one-shot metadata push to a live panel inside Resolve itself.

🎬

Send to Resolve

One click pushes descriptions, keywords, people, moods and picks into the Media Pool β€” with clips organised into bins, colour-coded by type, and flagged pick/reject. Smart Bins then sort themselves.

Free & Studio
🎨

Marker colours

Map metadata to Resolve marker colours β€” Yellow soundbites, Blue fact-checks, speaker turns, settled takes β€” laid onto every clip on send, to your own convention.

Soundbites Fact-check Settled
🧩

In-Resolve panel

A Workflow Integration panel inside Resolve: search footage, read transcripts & soundbites, push selects to a bin and drop markers β€” without switching windows. Loopback-only, opt-in.

No window-switching

Plus universal handoff β€” FCPXML timelines (Resolve & Premiere), Avid ALE, a MAM/archive CSV, and a producer-ready Story Bible in PDF or Word.

Yours, and only yours

Trust is the feature.

A documentary lives or dies on its rushes. HeyMonkey treats them β€” and your accounts β€” accordingly.

πŸ”’

Read-only ingest

Originals are read in place and never modified, moved or renamed. Proxies and data live in HeyMonkey's own cache β€” your camera-card structure is preserved.

πŸ”‘

Your own keys

You bring your own AI and transcription logins, stored encrypted on your machine. Nothing is shared, and no credits are ever charged without your say-so.

πŸ—„οΈ

Your own database

Every user gets a fresh, private library. Your footage, transcripts and notes never mingle with anyone else's.

πŸ“΄

Works offline

Local transcription and analysis run without the cloud. The Resolve bridge is loopback-only and off by default β€” nothing leaves the machine.

Stop logging. Start cutting.

Hand Resolve a project that already knows itself.

Point HeyMonkey at your footage and come back to a Library that's watched, transcribed, sorted and ready to edit.

Limited beta Β· runs on your own machine Β· bring your own AI keys

🐡
HeyMonkey
Ook ook! Welcome 🍌
Limited Windows beta

Request beta access.

We onboard testers in small batches on Windows. If you're a fit, we'll email your login. It runs on your own machine β€” footage never uploaded, bring your own AI key or go fully offline.

πŸͺŸ Windows beta Β· a Mac version is on the way

Request beta access

Tell us a little about yourself β€” no account is created yet, this just puts you on the list.

No account is created yet β€” this just puts you on the list. We'll never share your details.

πŸ΅πŸŽ‰

You're on the list!

Thanks β€” we've got your request. If you're a fit for the current beta batch, we'll email your login details soon.