We got tired oftimelines.
Every editor we'd used asked the same question first: where do you want to start? We thought that was the wrong question. The right one is what do you want — and then a team figures out where to start.
We spent years editing video the normal way — drag a clip, trim it, drag another, find the silence, cut it, write a caption, format it for four platforms, repeat. None of it was hard. All of it was slow, and none of it scaled past the hours in a day.
So we asked a different question: what if the editor wasn't one assistant, but a small team — each member good at exactly one thing, working in parallel, coordinated by someone whose only job is the plan?
That's EditorBuddy. You describe what you want in a sentence. An orchestrator reads your footage and your prompt, and plans a job: transcribe first, then clean up and analyze in parallel, then caption, then cut clips, then publish. Twelve specialists, each one narrow enough to be excellent, working off one shared plan.
The part that surprised us most wasn't the speed — though a 40-minute podcast becoming eight captioned shorts before your coffee is done is still strange to watch. It was that the team got better at knowing you. Caption styles, pacing preferences, the words you never want transcribed — all of it carries forward, project to project, the way a real editor who's worked with you for months would remember.
We're building this because we want to use it ourselves. Every agent on the roster below exists because one of us was tired of doing that part by hand.
The team, in full
Twelve agents, each with one job. The dot shows the state you'd see if you watched them work right now.
Orchestrator
planningorchestrator
Reads your prompt and plans the whole job as a graph — what runs first, what runs in parallel, what waits.
Memory
memorymemory_agent
Learns your caption style, pacing, banned words and voice. Brings them to every new project.
Transcription
workingtranscription_agent
Turns speech into a word-level transcript in seconds. Every other agent reads its work first.
Cleanup
workingcleanup_agent
Cuts silences, filler words and bad takes, then cleans the audio while it’s in there.
Subtitles
donesubtitle_agent
Captions in 48 languages, timed to the word and styled to your brand — not just legible, readable.
Clips
workingshorts_agent
Finds the moments worth posting and scores each one for hook, flow, engagement and trend.
Translation
planningtranslation_agent
Rewrites your script in another language without flattening the way you actually talk.
Dubbing
workingdubbing_agent
Speaks the translation in a voice that sounds like you, synced to the picture.
Editing
workingediting_agent
Auto-zoom, transitions, pacing. The cuts you would make yourself if you had the afternoon.
B-roll
idlebroll_agent
Pulls images, GIFs and footage to cover the stretches where nothing on screen moves.
Audio
idleaudio_agent
Music and sound effects, mixed under your voice — never over it.
Publishing
donepublishing_agent
Descriptions, hook titles, thumbnails, scheduling. Ships the finished cut to every platform.
What we keep coming back to
Plan before you cut
An orchestrator reads the whole job before any agent touches your footage. Fewer wasted renders, fewer surprises in the export.
No black box
Every agent's work is visible while it happens. You can watch the plan form, see what each one decided, and step in any time.
It should remember you
Your style — pacing, captions, banned words — carries from project to project. The tenth video should take less setup than the first.