Send.
Dump the camera roll straight into the chat. One clip or a whole album, plus photos, a voice note, a song, a link. No export screen, no upload bar, no file-size problem.
Bloom is the video editor in your messenger app. Send the footage, say what you want the way you'd say it to a person, and the finished video comes back cut, captioned and scored while you're off doing something else.
Bloom does all four on every video, before you ask for anything at all. Everything further down this page is extra.
Long pauses, fumbles and every “um” taken out, so the video moves. It's the difference between a take and a video, and it's the part that eats your evening.
Timed to the word rather than guessed, in a font that suits the footage, with the words that carry the line made to land. Not a default caption track dropped on top.
A small zoom, a cutaway, a photo. Some kind of pattern interrupt every three to five seconds so nobody scrolls past. Never obnoxious, and always following what you're actually saying.
It listens to tracks, picks one that fits the take, and tucks it under you so people hear you and not the song. Sound effects land where they punctuate.
Hand over the video, say how you want it, get it back. Exactly how it works with an editor, minus the hand-off, the waiting and the follow-up message asking where it got to.
There's nothing to learn here and no right way to phrase it. “Make it feel less boring” is a real instruction. Bloom watches the footage and comes back with the specifics.
Bloom watches every frame and reads the transcript before it suggests anything, then tells you what it's going to do in plain words. Nothing starts until you say go, and nothing is final until you say so.
final take, it rambles a bit
Here's the plan: down to 38 seconds on your three strongest lines, every pause gone, captions in your usual font with the numbers popped, a small zoom every few seconds, and a track tucked under you. Go?
go 🌸

captions a bit bigger, and lose the zoom at the start
On it. Same video, back shortly.
So change it. You just do it by replying instead of by dragging. Captions bigger. Lose the zoom at the start. That cut is early. Different song. It keeps working on the same video until you're happy with it.
What you give up is putting a caption on an exact pixel. What you get back is the hour and a half you were going to spend doing it, plus a video that doesn't lose your text overlays because you hit undo one too many times.
It will, because you tell it once. Captions run big. Never zoom on me. This font, this pacing, this kind of hook. It carries that into everything you make afterwards.
The rest it picks up from what you approve and what you send back, so each first cut lands closer than the last. Your format, your face, your voice. Bloom just does the assembly.
Most videos never need any of it. But the day one does, needing it is the whole problem, and it's the same conversation, not a different tool you have to go and learn.
Send the article, the dashboard, the product page. It goes and grabs the shot itself and puts it on screen at the moment you mention it, so you never stop to capture one.
You over a wall of images, picture-in-picture, the two-up and four-up formats. It cuts you out and builds the layout around what you're saying.
Cutaways matched to the line they sit under, not dropped in on a timer. Stock, YouTube or your own footage, and only where it's actually motivated.
Room noise, hiss and echo stripped out of the take you already shot, so you don't have to film it again in a quieter room.
Warm, cool, filmic or high contrast, with grain when the footage suits it. A proper grade, not a filter dropped over the top.
Remove someone who walked through the shot. Speed-ramp a reveal and freeze the reaction. Restyle a moment entirely. Rarely the job, but there when it is.
Ask for options and it makes them all: a punchier cut, a slower one, different music. You choose between finished videos instead of trying to describe the one you meant.
Send a rambling ten-minute take and get the strongest moments back as separate videos, each one cut, captioned and ready to post on its own.
Send the link to the video you took inspiration from. It watches that one, works out how it's cut, paced and captioned, then gives yours the same treatment.
Hooks, titles and stickers animated to your actual footage and your actual script, not pulled from a template pack. Most AI editors can't do this at all.
Dump the camera roll straight into the chat. One clip or a whole album, plus photos, a voice note, a song, a link. No export screen, no upload bar, no file-size problem.
Tell it what you want the way you'd tell a person. Or say nothing, and it will watch the footage, read the transcript and pitch you a plan first. Nothing starts until you say go.
The finished video lands back in the thread while you're doing something else. Not right yet? Reply and it keeps working on the same video.
Usually inside half an hour from the moment you approve the plan, and under an hour counting all the way back to when you first sent the footage. A long one with a lot of B-roll takes longer. The point isn't that it's instant. It's that it happens while you're doing something else, instead of eating the one to three hours it takes to do yourself.
There's no timeline to open and no clips to drag. You change it by replying: “captions bigger”, “lose the zoom at the start”, “the cut at twelve seconds is early”. It works on the same video until it's right. If you're particular about a single frame, say so in words. That's the whole interface.
Tell it once how you like things. Captions run big, no zooms, this font. It carries that into everything after, and learns the rest from what you approve and what you send back, so each first cut lands closer to your style. It's your footage, your voice and your face throughout.
Videos, one clip or a whole album at once, plus photos, voice notes, music and audio files, links, and plain text. Mix them freely in one thread and it keeps track of everything as one project.
No. “Make it feel less boring” is a valid instruction. Bloom watches your footage and reads the transcript before it suggests anything, then proposes the specifics. You just react.
You get a bit free to see what it does, no card needed. After that it's a subscription, and you can cancel any time from inside the chat. The plans are further down this page. Bloom is still early, so what a plan includes may shift as we work out what's fair.
Telegram, Discord, iMessage and WhatsApp today. You talk to it in a chat thread like you'd talk to anyone else. More messengers are on the way.
Your originals stay where you sent them and we fetch them when needed. Converted footage and anything Bloom generates for your edit is kept in our own storage so you can come back to a project, until you ask us to delete it, which you can do at any time. There's no public library, no feed, and nothing is published anywhere by Bloom. The privacy policy has the full detail.
Your first videos are free, no card needed. When you are ready for a plan, you pick one inside the chat, and you can change or cancel it there just as easily.
For creators still finding their rhythm.
For creators building for the long run.
Start free, in a chat
You filmed it, you wrote it, you showed up on camera. Hand the rest over.