Enes Akar
@enesakar
Followers
1,797
Following
781
Media
48
Statuses
2,146
Cofounder @upstash Previously @confluentinc and @Hazelcast
San Jose
Joined April 2008
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
#FreeJaprado
• 307700 Tweets
MINI
• 235502 Tweets
#SpanishGP
• 121043 Tweets
#光る君へ
• 114691 Tweets
Lewis
• 81208 Tweets
ブルアカ
• 71941 Tweets
#千鳥の鬼レンチャン
• 62904 Tweets
Lando
• 59556 Tweets
Norris
• 54384 Tweets
Chanyeol
• 50500 Tweets
Russell
• 39009 Tweets
Hamilton
• 38734 Tweets
藤牧京介
• 35427 Tweets
#ナイス投票
• 34434 Tweets
Verstappen
• 24358 Tweets
Sainz
• 23323 Tweets
水着カンナ
• 21764 Tweets
女子決勝
• 21114 Tweets
京ちゃん
• 21111 Tweets
ZETA
• 18726 Tweets
I MISS YOU GULF🩷
• 17996 Tweets
ヴァルキューレ
• 11378 Tweets
パーティクル
• 10701 Tweets
ネムガキ
• 10625 Tweets
We launched Serverless Kafka!
Now, you can have a fully managed Kafka cluster without paying hundreds of dollars.
🎊 Upstash Launches Serverless Kafka!
💯 Per request pricing with a monthly cap.
💯 Support Apache Kafka protocol.
💯 REST API for serverless and edge functions.
💯 Multizone replication.
💯 No charge on bandwidth.
💯 Free tier, no credit card required.
4
18
81
4
4
56
With so many recent releases and updates from us, I took the time to reflect and write about the story of Upstash. We started as just three people about three years ago, and since then, a lot has changed - you can read about the full story here:
4
6
55
We launched hazelcast cloud for beta. You will be able to setup and scale your cluster in seconds. It has free tier and $50 credits. Please try and share feedback.
0
30
43
I have developed a Chrome extension which uses ChatGPT API to improve your writing.
Select a text, ChatGPT suggests its better version.
Used
@Vercel
Edge for backend,
@Upstash
for rate limiting.
Thanks to
@nutlope
and
@steventey
for the inspiration.
5
7
40
Semantic search with Upstash
1
2
29
We've released the most requested feature for Upstash Vector. Now, you can perform SQL queries on vector metadata.
Our vector database just got a big upgrade - announcing metadata filtering for Upstash Vector!
🔸Query your vectors by metadata
🔸High-performance
🔸Intuitive, SQL like syntax
How it works & examples 👇🏻
4
4
56
2
3
28
Finally 😮💨
Now, you can use
@upstash
/ratelimit for your AI apps. It will update the rate by tokens not requests.
Thanks to
@_sourabpramanik
for the contribution.
💚 OS
Update for Rate Limit SDK! ⏳
Two new features:
◆ Custom consumption rates with the new rate parameter! Now, you can precisely limit requests based on the number of tokens. Great for rate limiting AI APIs that are priced by tokens rather than requests.
◆ New API:
0
4
29
1
3
25
I am working with computer science students on open-source projects. Comparing them with myself 20 years ago, I can easily say their productivity is 83 times higher.
1
0
24
🔥I was surprised that other vector databases are not doing this.
Introducing Upstash Embedding Models!
Starting today, you can insert and query raw string data with your Upstash Vector database.
◆ no more third party needed to vectorize text
◆ using high-quality open-source models
👇Detailed summary
6
13
100
4
0
24
I cannot find motivation to work on anything in a world where starving people seeking aid are bombed.
2
0
23
A startup is facing challenges and making tough decisions. It's disheartening to see some opportunistic competitors acting like vultures in these difficult times.
2
2
22
we are losing revenue but gaining many more users thanks to this
@t3dotgg
Not to mention it already uses in-memory cache under heavy load. So you wont rake up the upstash quota
0
0
4
0
1
21
AWS just announced AWS MSK (Serverless).
How can a service be serverless if it charges per hour?
#awsreinvent2021
6
5
19
Database-as-a-service companies create their own pricing dimensions (read units and write units). They guarantee their profitability over infrastructure costs, but pricing gets very complicated for users. Hard to understand, hard to calculate.
6
1
19
Serverless functions are great to run your backend. As they are stateless, you need an external data store.
So let's talk about what makes a database great for serverless.
👇👇👇
1
6
16
0
2
15
.. and when we implement query/search support on top of this, you may not need any other database.
🚀 Introducing Upstash JSON for Redis®.
As of today, store, retrieve and manipulate your JSON data with ease using Redis® commands. Full compatibility with RedisJSON. Check it out! 🔥
0
7
35
1
2
15
use
@upstash
/semantic-cache to cache your llm prompts. open source.
We're excited to introduce Upstash Semantic Cache, an open-source tool to easily:
◆ Classify natural language
◆ Cache your AI responses
◆ Make your API faster
👇 30-second demo
2
10
127
1
2
14
my second video, what makes upstash different?
1
4
13
upsy is happy today.
1
3
11
boolean canTestForCoronaInUS() {
return ( hasFever && hasCough && hasTravelled ) || isCelebrity || hasPower;
}
2
2
11
today i tried flowise. create a RAG chatbot just with drag-drop. it was smoother than i expected. impressed.
@FlowiseAI
1
2
10
Nowadays, it is hard say something against AWS. Half of the people tweeting about cloud is either AWS employee, or hero or champion or community builder or fanboy blah blah.
2
1
10
@mehulmpt
Yes. We aimed a fully multi tenant and serverless friendly offering. One requirement was to minimize the resource overhead while the database is idle. This was very hard with Redis. We decided to write it from scratch and design it for serverless from day 1.
1
0
10
Dear
@Apple
🍎 I know you're feeling the heat to keep up with the AI craze, but please don't just jump on the bandwagon. Don't create another LLM that won't be useful beyond impressing fanboys. You have something that all those other AI companies don't: you touch the lives of
0
0
8
@sandro_vol
@upstash
@momentohq
Hi Sandro. This will not affect us (
@upstash
). We do not use Redis, we implemented a KV compatible with Redis API from scratch.
The license restrictions on Redis will not impact the features of any Upstash product. `Upstash for Redis` is a custom key-value store implementation compatible with the Redis API. We implemented everything from scratch without using any code from Redis open source projects.
Our
27
52
635
2
0
9
`Sign up` links are big and everywhere. `Log in` links are small and hidden.
Existing users are not valued.
0
2
9
ok bullmq users, here's a plan just like you wanted. now only on fly, depending on its success, we can have it on aws too.
Introducing a new $10/month plan for Upstash Redis on . This plan is designed to optimize costs for BullMQ, Celery, and Sidekiq users with these limits:
> Single region
> Max 200MB storage
> Max 100 commands per second
Details:
2
1
22
0
1
8
namespaces -> multi tenant vector db
Introducing Namespaces for Upstash Vector! 🎉
Namespaces are like folders for your data. Instead of querying your entire database, namespaces allow you to query only parts of it.
For example, only data related to a specific user.
90-second demo 👇
3
8
72
0
2
8
I cannot stop myself from using llama-3 instead of gpt4. it's much faster, straightforward, more concise, and uses less fancy words.
0
0
8
Twitter search terribly sucks! I searched for "Josh" Josh (
@joshtriedcoding
) follows me, and I follow him. Twitter suggested every Josh except the one I'm looking for.
@elonmusk
, drop everything and fix the search.
3
0
8
Mini tutorial 👉 What is RAG? 👇
When building reliable and factually accurate AI tools, you’ll often hear the term “RAG” thrown around.
It’s a fancy way to describe a simple but very clever architecture to build AI apps - here’s how it works 👇
1
5
36
0
0
8
I used Google Cloud Run and Cloud Functions for small prototypes. Here my first impressions:
1
5
8
claude opus -> smart and to the point. my favorite so far
llama-3 -> very quick but too brief. maybe introvert
chatgpt4 -> talks too much, lacks directness. not my type
2
0
7
she is really good 💯
So very excited for the promotion of
@JenniferHli
to GP. She's a primary reason
@a16z
enterprise investing is what it is, being a key player behind so many deals, our team and our investment philosophy and process. She really is "the best of us".
70
10
391
0
0
7
i play bullet at to keep my brain active while chatgpt is writing my code.
0
0
8
I feel nervous to say something against AWS. Half of the cloud community is either AWS employee or AWS champion or AWS hero or AWS community builder...
4
0
7
My first coding video on youtube. It was harder than I expected.
Build a Serverless API with AWS Lambda, Upstash Redis and Serverless Framework
2
1
7
I was expecting a bigger gap between upstash and others. We will check the benchmark code.
0
1
7
if you are using upstash vector, you dont need to learn this
0
1
7
QStash enables long-running APIs for time-sensitive runtimes. It makes serverless use cases possible for non-serverless runtimes.
In the article below, the author wants to call a Vercel app from Slack Bolt (an SDK for Slack bots). The challenge is that Slack Bolt has a very
0
3
7
this is quite a long article to explain the pricing
0
0
6
success
@upstash
QStash is simply wonderful for background jobs!
Up and running in a handful of minutes -- on a Sunday evening 😁
0
2
7
0
0
6
Big day for QStash 🤩 Queue and batching support.
These were the most requested features from the users.
We’re happy to introduce 3 new QStash features! These suggestions come straight from you guys in the community.
◆ Ordered message delivery
◆ Parallelism - process multiple items in parallel
◆ Batching - batch multiple QStash messages
Everything new in 3 minutes 👇
4
5
81
0
2
6
Yeni projemizde birlikte heyecanlanacak, spring/docker/kubernetes sever arkadaslar ariyoruz.
0
7
6
@atehrosius
@alexbdebrie
Hey
@alexbdebrie
. The numbers you referred is from Upstash global database. For caching, you can select a single region database and when it is the same region with your application then latency will be <1ms which is good for caching usecase.
1
0
6
@cramforce
@upstash
@JericoPulvera
@steventey
@pbteja1998
@adam_janis
@dubdotsh
and ratelimit library leverages the runtime cache when possible. so the real cost will be much less than calculated. check here:
1
0
6
Another good use case of QStash.
New blog!
How to use QStash to decouple webhook data processing for guaranteed delivery.
by
@rubenpdegroot
1
3
49
2
1
5
last week, i have implemented my first rag app. the experience was smooth only after i use
@LangChainAI
instead of calling openai and pinecone apis myself.
3
0
6
🤔 20 minutes? that's longer than I expected.
@pwang_szn
This is great advice.
I encountered this when started implementing rate limits using Redis, had some issues with implementation then decided to switch to
@upstash
, took 20 minutes to set up everything without any problems 😁
0
2
4
0
0
6
REST is simple and clean.
0
2
3
0
2
6
