MTS @ context.dev classifier.dev. Scout @trygravityai & investor. Most startup problems can be solved w/ 100M views.
San Francisco
- Tweets 8,813
- Following 6,307
- Followers 15,846
- Likes 20,263
btw classifier.dev can classify images, video, and audio
literally my dream to have a product like this my entire life
classifier.dev
Zero-shot text classification over plain HTTP. No API key, no account.
classifier.dev
jev can now direct and make movies!
i built a live, audience-directed, choose your own adventure movie platform where the audience gets to comment the next scene they want to see
jev classifies each comment and picks the most relevant category to insert into the next prompt
one-shotted this app with GPT-6 and opus 5.5, with videos generated on hedra using minimax h3 max
i'm going to livestream this movie in the next hour or so but feel free to watch it in the link below 👀
5
10
1,239
i am once pleading you, i am begging you
give your agent everything it needs to do marketing
context dev for company, people, and bot-protected website data
classifier dev to classify hundreds of thousands of images and create simulations of thousands of users to test ad creative
databuddy to know which agents are using your product and how
elevenlabs to generate audio for x launch videos
i literally put some sweet api keys in the personalized section of codex because i can’t be bothered with skills/mcps
i am once pleading with you, i am begging you
you need to be giving your claude code all the tools it needs to do marketing
so that it can:
- make seed dance 2.5 ads with higgsfield
- do waterfall email enrichments with findymail, lead magic, apollo
- research the blog post it should write for x keyword using serper .dev
- have it scrape creators profiles with apify
- have it research links to build via data for SEO api
when you give it all of these tools
it is now your entire growth org
so get all these subscriptions today and give them to your agent
or get all of this here now - graphed.com/tools
3
1
13
1,125
i have a massive problem right now where agents are spending a shitton of credits dealing with code review/ci
it's right at the end of the session, so context is super full.
and then it needs to go through ci, read code review, fix all bugs, go through code review again, go through ci, etc.
this is costing a shitton of time and money.
has anyone solved this?
161
1
238
43,668
classifier.dev now has INFINITE context
classify documents of any length. go nuts
8
6
144
9,073
Dominik is one of the greatest product builders of our time
For the last month @miaugladiator1 and I have been obsessing over building the best platforms for companies to rank in AI Search
Today we are finally launching Notra GEO, the first step of our plan to achieve this!
2
21
2,009
claude's webpage tool is so frigging bad and so is codex's.
it's actively disgusting and messing up your agents
3
11
761
if anyone needs help with fraud detection, dm me.
have been dealing with it with classifier.dev and happy to advise
got fraud down 99.5%+ without banning entire countries
6
3
59
4,935
if you have fire office space somewhere around here for a team of 5-10 please dm me!
accelerating like crazy need a place to grow into
5
3
31
5,695
give your agents this i beg you
Introducing Scrape Anything.
we merged all our scraping APIs into one
inputs: websites, PDFs, DOCX, PPT, XLSX, 40 file types, anything.
into HTML, markdown, screenshots, images, JSON & more
Give us a URL, get your data
No need to worry about bot blocking, js rendering, proxies, iframes, shadow doms, retries, etc..
All the weird edge cases are our problem :)
3
25
831
128,933
before fraud protection I was spending $15/hour on classifier.dev
after < $1/hour!
if you suddenly found yourself unable to use it dm me, but this should massively improve usability for all legitimate free anonymous users.
guys, if you need classification, just make an account and pay.
it's dirt cheap. maybe even cheaper than the global proxy network your using to turn my savings into classification tokens.
3
28
2,532
Introducing Hotdog-27B, a specialized bipartite classifier.
Hotdog-27B beats jev on our independent binary classification benchmark.
I expect long term binary classification will have its own set of specialized open models, served on classifier.dev.
A significant percentage of global tasks are just 0/1 classification.
Is there a bug here?
Is this a hotdog?
Is this document relevant?
Is this coding trajectory valuable?
The real question is whether by focusing on just binary classification, we can meaningfully improve either accuracy, latency, or throughput.
All training code and benchmarks are public to encourage progress in this area:
huggingface.co/opensporks/ho…
github.com/mrmps/tinker-bina…
11
3
76
3,328
Michael retweeted
We use classifier.dev btw
classifier.dev
Zero-shot text classification over plain HTTP. No API key, no account.
classifier.dev 1
1
1
234
Hey guys.
Quick update: classifier.dev now serves several million requests a day, for free
I want to support images, video, and long-context on Classifier
But can’t without GPUs.
Already spending over $250/day on inference, and getting GPUs would put me in the red.
Anyone have extra compute I can borrow for a bit? Would help a ton.
I’ll put you on the website as a sponsor for the duration that your compute is used!
classifier.dev
Zero-shot text classification over plain HTTP. No API key, no account.
classifier.dev 26
18
236
22,748
