Loading video...

Video Failed to Load

Go Home

Introducing Shortcut v0.6 We released the first Excel AI agent 7 months ago and created a tidal wave Watch 4 messy, real-world examples that no other agent can do accurately The difference between a demo, and a product that top spreadsheet users at 1000's of companies love

237,332 views • 8 months ago •via X (Twitter)

54 Comments

nico's profile picture
nico8 months ago

Accuracy (1/2): Shortcut outperforms competitors, from the large labs to the smaller startups that have followed, on the messiest real-world tasks. These examples require work on large and complex models, extracting data from challenging attachments, etc.

nico's profile picture
nico8 months ago

Accuracy (2/3) Most Excel work starts with input data. Others struggle with large files, PDFs with embedded images of charts/graphs -- like the Roblox 10-K (where we score 100%), etc. Extract exact numbers from the worst documents

nico's profile picture
nico8 months ago

Accuracy (3/3) Excel AI is not meant to one-shot your work -- you will work alongside it. For any single prompt, you might not notice a dramatic difference between most products. The difference compounds fast over an hour, and this is deep agent + product engineering work

nico's profile picture
nico8 months ago

Audit-ability v0.6 will now auto open source PDFs and highlight the exact text to audit It's not enough to have the strongest accuracy; we need to give you tools to supervise AI. Rollbacks, multi-agent, multi-chat, comment notes, etc. are all great but they are table stakes

nico's profile picture
nico8 months ago

Work with files up to 250MB - Attach entire ZIPs of data rooms - Any file type... shortcut can even compress and edit videos Other agents have a 30MB limit This means real files, like general ledgers wont work Some startups don't even allow excel attachments..

nico's profile picture
nico8 months ago

Create almost any file with Shortcut Shortcut can now take your model and your attachments and build PDFs, PPTs, Word Docs... even GIFs Take a full data room, build a model, and turn it into an Investment Memo

nico's profile picture
nico8 months ago

Shortcut works across our Excel Add-in, Google Sheets Extension, our Website, and now our Beta API and Beta Desktop App Comment for Desktop & API access (we are rolling out!)

nico's profile picture
nico8 months ago

Full details on v0.6

nico's profile picture
nico8 months ago

Shortcut v0.6 is live at A moment is coming soon, if not already, where you have to be pretty crazy to not use AI for your spreadsheet work (Excel add-in available on our website too)

Jon Lai's profile picture
Jon Lai8 months ago

The best spreadsheet agent just keeps getting better!

nico's profile picture
nico8 months ago

The team is cooking!

Richard Pham's profile picture
Richard Pham8 months ago

Continuing to be first movers. Now we go

Finance Guy's profile picture
Finance Guy8 months ago

Is this AGI?

nico's profile picture
nico8 months ago

im feeling it

Shortcut's profile picture
Shortcut8 months ago

Can't wait to see someone edit a video on Shortcut ..

Dhaval Makwana's profile picture
Dhaval Makwana8 months ago

Seems cool

nico's profile picture
nico8 months ago

let us know how you find it!

Adrian Duermael's profile picture
Adrian Duermael8 months ago

Who needs 1.0 when 0.6 is this good? 😄

nico's profile picture
nico8 months ago

haha don't tease them!

toad's profile picture
toad8 months ago

Very cool. I (/ my analysts) struggle a little with model size and context. We have a beast of a model that is 80 tabs. Revenue reconciliations, really detailed P&L build up across 5 different reporting entities. What’s the best way to help rationalise and analyse the model?

nico's profile picture
nico8 months ago

Wow i have not seen 80 tabs before... I would love to put it to the test. DMing you

Ravi Shah's profile picture
Ravi Shah8 months ago

This is the excel ai that Microsoft & other companies want to ship, but can’t. Been using it for a few months, it keeps getting better

nico's profile picture
nico8 months ago

Glad you are loving it! Heads down on improving it for you

Frankie Li's profile picture
Frankie Li8 months ago

This is insane

Robert Yang's profile picture
Robert Yang8 months ago

Incredible

King.Khari's profile picture
King.Khari8 months ago

Genuine question... in what way is this better than Claudes excel add in? Why choose this over Claude.

nico's profile picture
nico8 months ago

Glad you asked!

King.Khari's profile picture
King.Khari8 months ago

Great stuff. Thanks

Peter Wang's profile picture
Peter Wang8 months ago

Difference between getting real work done and a flash in the pan

Hendrik's profile picture
Hendrik8 months ago

I’m one of your paid subscribers and am torn between continuing yours and Claude for excel. But I’m feeling the AGI towards v1.0. Keep the subscription at the moment.

nico's profile picture
nico8 months ago

Glad to hear! I hope you use both a bunch, make the best decision for you

Go, See, Do Awesome's profile picture
Go, See, Do Awesome8 months ago

Looks impressive, but I’ve yet to find an AI that can repeatedly with accuracy extract columnar data from an image or scanned page. Even PDFs often go weird. I’d be interested to know your approach to solving this issue or at least some stat on accuracy.

Justin Kaplan's profile picture
Justin Kaplan8 months ago

Absolutely sick great work

nico's profile picture
nico8 months ago

thank you sir!

Rai's profile picture
Rai8 months ago

Amazing Nico!

gazingback's profile picture
gazingback8 months ago

sick

nico's profile picture
nico8 months ago

thanks man

Mr K's profile picture
Mr K8 months ago

Wow, this is getting better faster than I expected - thought we’d be here by mid 27 - impressive

nico's profile picture
nico8 months ago

We are already at a point where people are falling behind by not using it, but by 2027 entire companies will be left in the dust

Mr K's profile picture
Mr K8 months ago

Agreed - I used to be able to plan 10- 20 years ahead, now 1 year seems impossible to plan - the next 2 years will be a shock for many people

neuronz - fidcs.com.br's profile picture
neuronz - fidcs.com.br8 months ago

I sincerely hope that you will be the company that wins this race in the automated spreadsheet market. Microsoft don't deserve.

researcher's profile picture
researcher8 months ago

congrats! Can i get api access please

nico's profile picture
nico8 months ago

yes dm me your email please!

Kushagro Bhattacharjee's profile picture
Kushagro Bhattacharjee8 months ago

Yesterday I learned that Shortcut can fully process my screen recordings

nico's profile picture
nico8 months ago

lol i had it compress a whole video for me too

Adele Smolansky's profile picture
Adele Smolansky8 months ago

Amazing work

Inflectiv AI ⧉'s profile picture
Inflectiv AI ⧉8 months ago

The team has really been shipping, keep the momentum going🔥

nico's profile picture
nico8 months ago

its all we know ❤️

incognito's profile picture
incognito8 months ago

What’s the usage limit on this? Just purchase Claude. will move over if the limits are better.

nico's profile picture
nico8 months ago

No one complains about limits. You can purchase credits and they drain much slower

Taylor Johnson's profile picture
Taylor Johnson8 months ago

Holy showing my dad this retirement calculator to see if he likes it more than annual pen and paper

nico's profile picture
nico8 months ago

Lmk 👀

Tom English's profile picture
Tom English8 months ago

With big AI labs entering the competition, I would suppose that you'd need to lean heavily on AIs in further development of Shortcut. Mind commenting on that? In any case, I'm rooting for you.

Hanif Dean's profile picture
Hanif Dean8 months ago

hey nico, i've been using this to build out a few models just for my personal equity investments - but i noticed that: 1. are workspaces constrained to single chat session? (because ideally whenever i open/re-open file, i'd want the note to directly link to the source) (cont)

Related Videos

Introducing LobeHub: Agent teammates that grow with you. LobeHub is the ultimate space for work and life: to find, build, and collaborate with agent teammates that grow with you. We’re building the world’s first and largest human–agent co-evolving network. Two years ago, we built LobeChat, an open-source interface for using different AI models. Today, LobeChat has 70k+ GitHub stars and serves 6M+ users worldwide. How to fully unlock the power of models has always been a shared mission between us and the community. We started with interaction — a fundamentally new, agent-first experience. Agents are no longer passive tools invoked in a single conversation. They should be proactive, always-on units of work. Treating agents as the minimal atomic unit is also the core of our agent harness infra. Today’s agents are mostly one-off executors. Even with memory, it’s often global — and hallucinates. We build long-term agent teammates that evolve with users. Each agent has its own dedicated memory space, editable by users, allowing humans and agents to co-evolve over time. This, in turn, allows us to design clearer rewards for reinforcement learning and create cleaner environments for continual learning. Agent teammates can work in groups. Through a multi-agent system, agent groups operate faster, more cost-effective, and go beyond what single-agent systems can achieve. For example, a single agent often requires heavy user involvement to proceed step by step, whereas LobeHub can execute the same work from a single instruction, with a supervisor orchestrating agents that run in parallel or debate to produce better results. We are building the collaboration network among agent teammates — and between humans and agent teammates as well. Ease of use matters. AI intelligence and shared human intelligence are equally important. With simple instructions and tool selection, you can effortlessly build and team up with agent coworkers to deliver complex, systematic work — even assembling a quant team to execute trades. Through the LobeHub community, anyone can discover, reuse, and remix agents and agent groups, customizing them to fit their own workflows, preferences, and needs. Last but not least, our vision started with LobeChat: multi-model support is the most efficient approach for users. We believe different models excel in different scenarios. By routing across multiple models, LobeHub improves cost efficiency and unlocks capabilities that a single-model setup cannot easily support.

LobeHub

185,462 views • 8 months ago