Graphify Labs (YC S26)

knowledge graph engine for your codebases
backed by @ycombinator

[Safi @safishamsii]

The best way to sell and improve a product is to use it yourself every day.

That's why every PR on @graphify OSS is reviewed by Graphify.

Not a self-assessment. Real code graph analysis for coupling regressions and blocking issues.

Competitive on SWR-bench.
Try the same reviewer before we launch→ https://t.co/0B1iwKRo8k

[Safi @safishamsii]

People see the stars. People see the downloads.

They never see the pace.
Build. Ship. Iterate. Repeat.

That’s the part that actually built @graphify

Repo: https://t.co/ANYBBTda8R

[Safi @safishamsii]

What a privilege it is to be overwhelmed by a life we once prayed for.

Got into @ycombinator on my first attempt as a solo founder with @graphify

Still doesn’t feel real.

[Safi @safishamsii]

Almost 100k downloads in a single day.

99.3k in the last 24 hours.
4.5M all-time.

New record for @graphify on PyPI.

Thank you to every contributor, every user, and everyone who shared it.

[Safi @safishamsii]

Sold the @graphify pro plan last night to a random guy from Oxford at WeWork

Had him connect MCP from our platform to Claude desktop

Then pull up his own repo and show the reply + token savings live

Best way to sell is make the product actually work

Token savings is just the side effect

[uv @yuv44n]

add graphify to the mix and you're saving a ton of tokens in a big enough codebase

[Safi @safishamsii]

Just been a week since I made @Graphify platform public. No launch. No announcement.

2,500+ people already signed up.

https://t.co/RKxsRdSv86

[Sean @seandotio]

I love using Graphify now for my coding agents. It's fantastic knowing they have a proper map of my codebase, and it's dependencies, and that they're not wasting a ton of valuable context window grep'ing around to try to understand the codebase on each fresh instance.

@graphify

https://t.co/i1s5PdwdxI

[Graphify Labs (YC S26) @graphify]

Code generation is solved.
Proving the code is correct is not.

Human defines the ontology and properties first.

LLM implements against them.
Then we formally verify the design.

Every single line of code needs a formal proof.

That’s where this has to go.

[Jared Friedman @snowmaker]

Graphify Labs is the first YC company to hit 100K github stars during the batch.

[Safi @safishamsii]

100,000 stars on @graphify
First Indian developer to ship a 100k-star repo.
Only the 5th @ycombinator -backed open-source project to ever hit 100k.

Stars are vanity until they aren’t.
100k is where they stop being vanity.

[Safi @safishamsii]

Go test https://t.co/RKxsRdT2XE until we release the V1

[Jay @jaykhan]

building the LegalOS company brain using @graphify

[Water flowing | anbello.eth | anbello.tez @Waterflowing0]

Preparing knowledge graph for leaf venation algorithm optimization with @graphify

[Safi @safishamsii]

300+ cracked devs in the @graphify Discord already.

If you’re building AI agents or knowledge graphs that need real memory, this is the spot.

https://t.co/fsAJYvNBPn

Tell us what are you shipping this week on the server!!

[Nick Shilov @gaussdevhere]

tried @graphify today, great experience

helped to get organized with all modules

@safishamsii you cook

[Lopez @LopezOnchain]

Thought it would be cool to see @graphify chart in 3D so I made it

[Safi @safishamsii]

Our Open source @graphify has gone past @bunjavascript to be top 6 Open source repositories backed by @ycombinator ever

https://t.co/ANYBBTdHYp

[rob. @robiartec]

LE DICES A TU IA QUE ANALICE TU PROYECTO Y EMPIEZA A RECORRER ARCHIVOS UNO POR UNO DURANTE MINUTOS.

Ya existe una forma mucho más inteligente de hacerlo. Y ya supera las 86.000 estrellas en GitHub.

Se llama Graphify. En lugar de leer cada archivo desde cero, crea un mapa completo de tu proyecto para que la IA navegue directamente por lo que necesita.

→ Indexa código, PDFs, imágenes y vídeos con un solo comando
→ Diferencia relaciones reales de las inferidas
→ Compatible con Claude Code, Cursor y más de 20 agentes de IA
→ Todo se ejecuta de forma local

En vez de hacer que la IA recorra medio repositorio cada vez que preguntas algo...
Encuentra el contexto exacto en segundos siguiendo las conexiones de tu código.

Cómo haces para que tu IA no se pierda cuando el proyecto empieza a crecer? Repoo abajo.

[Safi @safishamsii]

Race is on to become the first Indian developer to ship a 100k-star repo

Also Currently on track with @graphify to become the 5th @ycombinator backed open-source project ever to hit 100k stars

https://t.co/ANYBBTda8R