Back to directory
AI & ML · AI video & image generation

Palette

Your entire studio in one prompt. From first idea to final output, keep the entire creative workflow in one conversation instead of stitching together separate tools.

building @trypalette_ai (YC S26)
San Francisco332 followers
TLVC Rating
Hook
Editing / Creativity
Copy
Sentiment of launch
Distribution strategy
Community Rating
No ratings yet
Your rating
Sign in to rate this launch.

About

Palette is a single-conversation platform for generating, editing, and directing video and image content, aimed at enterprise creative teams and independent content creators who currently juggle separate tools for storyboarding, generation, editing, and post. From one prompt or reference sheet, it produces scenes that stay consistent across shots, characters, and styles, and lets users revise the output in plain language, swapping objects, restyling scenes, or recutting a long take into a tighter edit without leaving the thread. Under the hood, it routes each subtask to the model best suited for it, sending lip sync to a voice model, physics-heavy shots to Kling, anime looks to a separate anime model, and b-roll to whichever provider is cheapest, so users get a coordinated pipeline rather than a single general-purpose generator. The company positions itself as an interactive multimodal generation platform that helps people create, personalize, and maintain multimedia content at scale , with target verticals spanning marketing, entertainment, learning and development, customer support, product, manufacturing, healthcare, retail, field service, and safety training. Character and brand consistency are core to the pitch, users define a character or reference sheet once and carry that visual language through multiple shots and formats, which matters for teams producing recurring ads, training modules, or episodic content where identity drift across clips is a common failure mode of current generative tools. The launch is notable because it lands as the market fragments into dozens of single-purpose generators, and Palette is betting that orchestration across those models, plus persistent context in one workspace, is more valuable than any individual model advance. The company was founded by Serena Pei, an MIT computer science undergraduate previously affiliated with the MIT Laboratory for Information and Decision Systems and MIT FutureTech, and the team describes itself as MIT computer science graduates and multimodal AI researchers. Palette is part of Y Combinator's S26 batch and lists Y Combinator and OpenAI among its backers on its site.
Tags
B2C<500KSeedProduct launchExplainerB2BGlobalAI-generatedDemoUSVertical AI
Comments (14)
Sign in to join the discussion.
Nikhil Raghavan11d ago

The 'entire studio in one prompt' bit is doing a lot of heavy lifting. Show me the timeline scrubbing before I believe it.

Marta Kovač11d ago

That launch video cut from ideation to final render in like 4 seconds and I refuse to accept that pacing without a blooper reel.

Priya Ganesh11d ago

Every demo like this looks flawless in the trailer and then falls apart when I try to make a 12 second product ad with two humans in frame. Prove me wrong.

Tomek Zieliński11d ago

The b-roll of Photoshop, Premiere, and Figma tabs closing one by one was chef's kiss editing. Whoever cut this deserves a raise.

Omar Diallo11d ago

ok wait, enterprise teams AND content creators in the same sentence is a bold positioning choice. Pick a lane or make sure the onboarding knows which door you walked through.

Lin H.11d ago

Tagline rewrite, on the house: 'One prompt in. Whole campaign out.' You can Venmo me a coffee.

Fatou Mbaye11d ago

Curious what the API surface looks like and if there's a docs site yet. Multimodal engines live or die by how fast I can hit /generate.

Yusuke Tanabe11d ago

MIT researchers plus YC S26 plus multimodal plus enterprise. That's a full buzzword bingo card, respect.

Juno Park11d ago

How many people on the team? I want to know how cooked I should feel.

Chidi Obi11d ago

Building in adjacent territory and I have to say the 'one conversation' framing is smart. Stealing it in spirit, not in copy.

Hanna Lindqvist11d ago

First-tap test: does opening Palette drop me into a blank prompt or a templated 'try this' screen? One of those retains, the other churns in 40 seconds.

Ravi Bhatt11d ago

Any chance you're open to chatting about the enterprise pilots? Would love to hear which verticals bit first.

Sana Iqbal11d ago

Impressions look healthy but the reply-to-like ratio is telling me people are watching quietly and not sure what to say yet. Classic 'is this real' phase.

Deva Krishnan11d ago

Real question for the infra folks: is that video generation happening in the conversation or are you queuing jobs and streaming previews? Big difference for pricing.