Videoclaw is a Mac desktop app that turns a chat prompt into a finished video, handling script drafting, voiceover generation, AI avatars, media sourcing, and editing through a single agent you talk to instead of a timeline you scrub. It targets founders, marketers, social media managers, and small teams who need launch videos, product demos, explainers, ads, podcast cuts, or short-form pieces adapted from an X post or newsletter, without touching a traditional editor. The workflow runs in five named stages, chat, script, generate, preview, and download, and the app ships with a library of templates covering formats like founder-led demos, kinetic-type documentaries, screen-recording walkthroughs, and talking-head reels.
The launch matters now because this is the move from private alpha to public beta. The company behind it is Humeo, a San Francisco applied AI lab focused on creative intelligence, and Videoclaw is its first shipped product. Co-founder Joseph Chin is building the team out of San Francisco, positioning Videoclaw as an AI-native video creation tool rather than a plug-in for an existing editor. The launch video was itself made entirely inside Videoclaw, which doubles as the clearest demo of what the agent can string together end to end.
One design choice worth flagging for operators evaluating the space. Videoclaw asks you to connect your own ChatGPT or Claude account for the reasoning layer, so you bring the model subscription you already pay for and Videoclaw supplies the tools, the media pipeline, and the orchestration around it. That makes it an orchestration layer rather than a model vendor, which is a different bet than most competitors shipping this month, and it means the product inherits whatever the underlying labs release next.
Comments (13)
Sign in to join the discussion.
N
Nadia K.22d ago
The demo video being made in Videoclaw is the flex, but you buried it in the last line. That should have been the hook, not the P.S.
S
svetlana22d ago
Cool demo but I've seen four of these this month and they all edit like a caffeinated intern. Show me a two minute cut that doesn't cut every 0.8 seconds.
T
Temi Adewale22d ago
Prompt to video is the new prompt to app is the new prompt to landing page. At what point do we just prompt the prompt?
K
Kentaro I.22d ago
Would love to chat about the origin story for a piece I'm working on. DMs open whenever you have a minute after launch day settles.
R
Raghav Pillai22d ago
How big is the team behind this? Because if it's under 5 people I'm going to quietly cry into my Notion doc.
M
Marco D.22d ago
API? Webhooks when a render finishes? Rate limits per workspace? Asking because I want to plug this into a pipeline, not click buttons like it's 2015.
H
Hiroko Tanabe22d ago
Question from someone new to this space: what happens when the prompt is vague? Does it just pick vibes for you or does it ask follow ups?
B
Benoît L.22d ago
This is one of the last video tools that will ship without an agent layer doing the prompting for the user. Mark my words, in six months you'll add it and call it Videoclaw 2.
P
Priya Sundaram22d ago
Reminds me of an early portfolio bet in the generative video lane. The winners here won't be the best model, they'll be the ones who nail the edit brain.
D
Dave M.22d ago
Nice launch. One thing I'd add to the roadmap: brand kits with locked fonts and color palettes. Every team will ask for it by week two.
Q
quinn22d ago
You have video ideas, Videoclaw makes them. My therapist has ideas too, Videoclaw doesn't make those.
E
Ellen Ørlund22d ago
Genuinely curious what's under the hood. Is this orchestrating existing models or is there a fine tuned editor doing shot selection?
Y
Yusra H.22d ago
I'll believe the 'just prompt' promise when I see the raw prompt AND the raw output side by side. Every demo I've seen this year has been curated within an inch of its life.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Videoclaw
You have video ideas. Videoclaw makes them
You have video ideas. Videoclaw makes them
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Videoclaw is a Mac desktop app that turns a chat prompt into a finished video, handling script drafting, voiceover generation, AI avatars, media sourcing, and editing through a single agent you talk to instead of a timeline you scrub. It targets founders, marketers, social media managers, and small teams who need launch videos, product demos, explainers, ads, podcast cuts, or short-form pieces adapted from an X post or newsletter, without touching a traditional editor. The workflow runs in five named stages, chat, script, generate, preview, and download, and the app ships with a library of templates covering formats like founder-led demos, kinetic-type documentaries, screen-recording walkthroughs, and talking-head reels. The launch matters now because this is the move from private alpha to public beta. The company behind it is Humeo, a San Francisco applied AI lab focused on creative intelligence, and Videoclaw is its first shipped product. Co-founder Joseph Chin is building the team out of San Francisco, positioning Videoclaw as an AI-native video creation tool rather than a plug-in for an existing editor. The launch video was itself made entirely inside Videoclaw, which doubles as the clearest demo of what the agent can string together end to end. One design choice worth flagging for operators evaluating the space. Videoclaw asks you to connect your own ChatGPT or Claude account for the reasoning layer, so you bring the model subscription you already pay for and Videoclaw supplies the tools, the media pipeline, and the orchestration around it. That makes it an orchestration layer rather than a model vendor, which is a different bet than most competitors shipping this month, and it means the product inherits whatever the underlying labs release next.
Comments (13)
The demo video being made in Videoclaw is the flex, but you buried it in the last line. That should have been the hook, not the P.S.
Cool demo but I've seen four of these this month and they all edit like a caffeinated intern. Show me a two minute cut that doesn't cut every 0.8 seconds.
Prompt to video is the new prompt to app is the new prompt to landing page. At what point do we just prompt the prompt?
Would love to chat about the origin story for a piece I'm working on. DMs open whenever you have a minute after launch day settles.
How big is the team behind this? Because if it's under 5 people I'm going to quietly cry into my Notion doc.
API? Webhooks when a render finishes? Rate limits per workspace? Asking because I want to plug this into a pipeline, not click buttons like it's 2015.
Question from someone new to this space: what happens when the prompt is vague? Does it just pick vibes for you or does it ask follow ups?
This is one of the last video tools that will ship without an agent layer doing the prompting for the user. Mark my words, in six months you'll add it and call it Videoclaw 2.
Reminds me of an early portfolio bet in the generative video lane. The winners here won't be the best model, they'll be the ones who nail the edit brain.
Nice launch. One thing I'd add to the roadmap: brand kits with locked fonts and color palettes. Every team will ask for it by week two.
You have video ideas, Videoclaw makes them. My therapist has ideas too, Videoclaw doesn't make those.
Genuinely curious what's under the hood. Is this orchestrating existing models or is there a fine tuned editor doing shot selection?
I'll believe the 'just prompt' promise when I see the raw prompt AND the raw output side by side. Every demo I've seen this year has been curated within an inch of its life.