Ehsan Aslani

Software Architect/Developer

[Whisper.Security @WhisperSecTech]

1/6
📢We just shipped Whisper AI Context.

A free-tier MCP server that gives your security AI agent live infrastructure intelligence, BGP, DNS, WHOIS, threat-graph, in 2 minutes of setup.

[60-second install screencast]
👉https://t.co/mb5HJq6DJE

🧵

[Ehsan Aslani @ehsan_a]

@msh_1486 @MohsenSeyedain ممنون

[Ehsan Aslani @ehsan_a]

@MohsenSeyedain 😂😂😂

[Ehsan Aslani @ehsan_a]

@MohsenSeyedain اون که صد در صد ولی این داستان سوبسید انرژی هم معضل بزرگیه که متاسفانه هیچ راه حلی نداره و متاسفانه به نظر من نجات اقتصاد ایران کلا امکانپذیر نیست دیگه.

[Mehran @mehrantsi]

Ultra-fast, embedded KV in pure Rust.
<300ns GET and <600ns INSERT.

[Mehran @mehrantsi]

Is it obvious when we switched to HPKV v1.18? The processing time for write and delete operations are ~100x faster!
P.S: These are sample end-to-end times, from the moment request is received to our API to when the response is being sent. This is our EU1 (Germany) server.

[Mehran @mehrantsi]

HPKV's memory MCP server is more than a simple save and retrieval, because models won't be able to search for exact keys that they once used. It allows models ask questions and get contextual answers. It provides a RAG based search as well as Q&A based context retrieval, so the model can always get the right answers.

[Mehran @mehrantsi]

I recently had some fun creating a search on 3.2 billion URLs from Common Crawl Index with <10µs lookup time and a 48€/month server.

https://t.co/5BS2ZFMCRb

[HPKV @hpkvHQ]

Nexus Search adds Retrieval Augmented Generation (RAG) capabilities to HPKV, enabling semantic search and AI-powered question answering over your key-value data.
https://t.co/u3fzBa0b3r

[HPKV @hpkvHQ]

Checkout this chat application created by leveraging HPKV's pub-sub feature and see how easy it is to create a realtime application. 🙌🏼

The only back-end implementation here is initial token generation! The rest is fully on front-end, leveraging HPKV's WebSocket API.

[Ehsan Aslani @ehsan_a]

Building a blazing fast chat application in a remarkably simple way.
You will need:
No websocket servers
No dedicated backend (beyond security token generation)
No message brokers
No additional databases for message persistence

https://t.co/jhEIKsilBm

[Mehran @mehrantsi]

Version 1.0.16 of hyperlimit released.

Simpler tenant based rate limiting.

https://t.co/eD6SoFDRPO

[Mehran @mehrantsi]

HyperLimit: Anative rate limiter for Node.js

Native C++, lock-free design, zero dependencies.

Framework specific middleware for Express, Fastify and Hyper-express or build your own using the core package.

MIT license btw.

[Mehran @mehrantsi]

1/ I've recently been working on a high performance, persisted storage system with sub-microsecond read & write latencies (250-650ns range). During my pursuit for raw performance, I went down into a few rabbit holes, one of which was free space management, used in filesystems. 🧵

[Ehsan Aslani @ehsan_a]

@mahenoghre آفرین به تو فاشیست نژادپرست آریایی

[Ehsan Aslani @ehsan_a]

@MohsenSeyedain وجه حماسی هم ممکنه پیدا کنه. مثل بهار عربی.

[Ehsan Aslani @ehsan_a]

@MohsenSeyedain @MaedehMN اون قراره هم تو رو یاد بگیره هم خودشو. 😁کلا یادگیری رو محول کن بهش. بعد هم فعلا عضو جدید جامعه است، البته ما اینجوری فکر می کنیم ولی نمی دونم اون ما رو عضوی از جامعه اش می دونه یا نه.

[Ehsan Aslani @ehsan_a]

@MohsenSeyedain نه نیست واقعا. ولی می دونی چیه، من هر دفعه میام ایران حس می کنم داره بیشتر شبیه بقیه دنیا می شه. یعنی از خاص بودن و ایران بودنش داره کم میشه.

[Ehsan Aslani @ehsan_a]

@MohsenSeyedain موافقم که برگشتی در کار نیست. مگر اینکه طرف دلش خیلی واسه خونه تنگ شده باشه. واقعیت اینه که جامعه هم‌ نیازی به اون افراد نداره و اومدنشون کمک زیادی هم نمی کنه.