Helicone @helicone_ai Twitter profile

Pinned Tweet

Helicone

2 months

📣📣Hello world, we are re-introducing Helicone! P.S. we spent a long time on the video, so sound on, pretty please? 🥺 Let us know what you think! Sign up for free → Get in touch →

17

25

77

Last Seen Profiles

@web3ascend

@hajime_moriyama

@_DVOUR

@DavidQuart38411

@ImClique_

@OUR00000000

@CoolBarrrr

@TCR_news_

@tokuspry

@Laurengraphics

@QueenReina2010

@danieljsongray

@anchorlang

@IvoBrandr

@ShinoAwoo

@ccabellopure

@Lucaes24

@msitewei

@Gaurav_212005

@GabeSm1th

@DeviousAngel_VT

@WaxingMoonlight

@BON_SBNation

@fthrihvsnnd

@stephanivictor

@NashvilleDeeDee

@vats9_9

@DoNotDisturbYan

@schupbach

@tismoonie

@cgdarreon

@sponcor

@q3pzIXsbtk66651

@SinemoreCooper

@AEHATT

@HotaruNara

Helicone

@helicone_ai

1 year

Does anyone else notice @OpenAI is a bit slower than last month? Well, we have been measuring its latency. Last week `gpt-3.5-turbo-0301` was 5x slower than it was in late April. This is based on over 10m data points

39

41

538

Helicone

@helicone_ai

10 months

🎉 Re-Introducing Helicone, the top Monitoring solution for your LLM calls 🎉 With only two lines of code you can log requests, monitor spending, track latency and so much more with Helicone. We're proud to be open source and would love more eyes on our Github below!👇

12

35

267

Helicone

@helicone_ai

1 year

We added caching support to Helicone Easily cache your OpenAI requests with our proxy so that duplicate requests don't drive up your bill. This also makes testing and developing a lot easier, faster, and cheaper!

10

100

Helicone

@helicone_ai

1 year

Announcing Helicone, your observability platform for GPT-3 Get detailed metrics for usage, latency, and costs with a single line of code

1

12

66

Helicone

@helicone_ai

1 month

🎉🎉 We’re super excited to announce that @helicone_ai has been selected as 1 of the 29 startups to participate in @Cloudflare Workers Launchpad this cohort! 🧊 At Helicone, we're on a mission to bring the most comprehensive insights to developers building in #GenAI . In just

6

4

37

Helicone

@helicone_ai

10 months

We're excited to announce we've logged 100 MILLION requests through Helicone! 🎊🎊

3

1

33

Helicone

@helicone_ai

4 months

🚨Last week, we saw @wenquai using @helicone_ai to fine-tune a model in order to lower his AI costs by over 99%. So over the weekend, we built out a fine-tuning feature that lets you do this all within Helicone. Here’s how:

4

7

27

Helicone

@helicone_ai

1 year

gpt-4

2

0

24

Helicone

@helicone_ai

6 months

Scared about OpenAI’s up time? Introducing 🚀Gateway Fallback🚀 We make it really easy to fallback to another provider.

Gateway Fallback (Beta) - Helicone

With the instability of some APIs and the similarity of many providers, we have developed a unified gateway. This gateway allows you to utilize any provider through a single endpoint.

docs.helicone.ai

2

3

25

Helicone

@helicone_ai

3 months

Fine tuning a model in less than a minute ft. @subwaysurfers !🛹🚆 (our intern made this) We want people to be able to quickly fine tune their model with with minimal code. (our intern wrote this) Comment +1 if our intern did well! (we'll buy him pizza if 10+ people reply)

11

3

23

Helicone

@helicone_ai

5 months

Working at a startup means you can make your landing page holiday themed 🎄🎅

1

2

23

Helicone

@helicone_ai

9 months

Good job @OpenAI on increasing capacity 👏, latency has definitely gone down a bunch However, errors are on the rise 👇 5XX/(5XX+200) status

1

2

22

Helicone

@helicone_ai

8 months

🚀 New at Helicone: User Feedback on Responses! 👍🏼👎🏼 Rate AI outputs, refine prompts, and drive model improvements. 📊 Track positive and negative feedback trends over time! Dive in: .

2

3

22

Helicone

@helicone_ai

1 month

🎉🎉 We're excited to announce that Helicone is now integrated with @togethercompute ! You can now run observability on Together with the most up-to-date pricing and stats - all with a custom domain: . We look forward to continuing to bring openness and

1

4

20

Helicone

@helicone_ai

13 days

We’ve teamed up with @posthog to bring your LLM analytics to your PostHog dashboard, so you can view them all in one place! 🤝 In classic Helicone style, you can do this by modifying just a few lines of code. Check out our step-by-step guide in the 🧵:

1

0

20

Helicone

@helicone_ai

3 months

🎂It's officially been a year since we started Helicone (originally Valyr)! 🎂 In one year we're proud to say we now: 🥳log 120,000,000 reqs per month 🤯have logged 1 BILLION requests 👩‍💻have over 6,000 users Thank you so much to our community and users!

4

0

19

Helicone

@helicone_ai

10 months

We want more people in AI to meet each other: Say hi 👋 so people see you in the comments!

20

3

18

Helicone

@helicone_ai

8 months

🔐 Introducing Vault, Helicone's solution to AI provider key management. With Vault, businesses can securely store and distribute provider API keys, like those from OpenAI. By handing out Helicone proxy keys, we guarantee all interactions are channeled through Helicone, ensuring

1

3

17

Helicone

@helicone_ai

4 months

@wenquai @wenquai hey Zachary! We're glad you're enjoying using @helicone_ai . We spent a lot of time on our integration and are glad it was an easy plug-and-play! We're also playing around with fine-tuning on our platform if you're interested 👀

1

14

Helicone

@helicone_ai

7 months

The Helicone team depicted by @OpenAI 's Dalle-3 @NguyenScott7 - @justinstorre - @coleywoleyyy

1

0

16

Helicone

@helicone_ai

9 months

🚀 We've hit 200 million requests! 🚀 We want to give a huge thank you to the Helicone community using our product to make building on LLMs easier and more efficient! Please tag us in any projects using Helicone and we'll retweet it!

0

3

16

Helicone

@helicone_ai

10 months

We are thrilled to announce our collaboration with @Auto_GPT , aimed at enhancing their benchmark logging system. Let's make great strides together! 💪

3

5

15

Helicone

@helicone_ai

6 months

🚀 We're excited to share our latest blog on Rosetta, our solution to a fragmented LLM landscape. Learn how we're achieving universal LLM support by standardizing LLM schemas with the power of generative AI! 🚀

2

4

16

Helicone

@helicone_ai

1 year

Excited to announce custom rate limits for your OpenAI requests to manage costs and traffic! Define a policy like 1000 requests per minute over any segment like users or custom properties (IP addresses, orgs, etc)

1

16

Helicone

@helicone_ai

1 year

Caching for LLMs is one of our most used features, so we released a dedicated page for it For any application with commonly asked prompts, it's a huge win for costs + latency!

2

3

15

Helicone

@helicone_ai

9 months

Open-Source enjoyers rise up

1

3

15

Helicone

@helicone_ai

2 months

🎉 We are excited to welcome Stefan to our engineering team! 🚀 Stefan has 6+ years of experience in engineering and fintech. Before Helicone, Stefan spent time building LLM apps at startups ranging from pre-seed to Series A+. Welcome Stefan! #NewTeamMember

2

15

Helicone

@helicone_ai

7 months

Keep track of your user's feedback in Helicone🎉 Easily capture and maintain feedback (👍/👎) within the Helicone dashboard. Learn more below 👇

1

14

Helicone

@helicone_ai

10 months

Developer using LLMs? 🧑‍💻 We're dropping something new... Reply 👋 for early access

17

1

14

Helicone

@helicone_ai

6 months

Helicone's first Hackathon co-hosting w/ @posthog ! Free food, prizes, and swag!

2

14

Helicone

@helicone_ai

7 months

Dall-E 3 is amazing! Which ones do you all like best?

6

1

14

Helicone

@helicone_ai

1 year

🎉 Excited to announce our partnership with Y Combinator! We're on a mission to make generative AI observability & monitoring a breeze for everyone building with it Here's what Helicone has achieved so far (just a sneak peek!) 🧵

Y Combinator

@ycombinator

1 year

Congrats on the launch and welcome to YC, team @helicone_ai ! Helicone (YC W23) is an open-source observability platform for generative AI companies to log all of their requests made to OpenAI and track their costs, result quality, and latency.

6

11

92

1

2

14

Helicone

@helicone_ai

1 year

Helicone is now available as Python and Node packages to make logging, caching, and custom rate limits for OpenAI easier than ever

3

0

12

Helicone

@helicone_ai

1 year

We ❤️ building on @vercel

Steven Tey

@steventey

1 year

8. ( @helicone_ai ) – Observability for LLM apps

1

0

25

1

0

13

Helicone

@helicone_ai

7 months

Introducing Key Vaults on Helicone🎉

0

3

13

Helicone

@helicone_ai

10 months

Check out our Github and leave a star!

GitHub - Helicone/helicone

Contribute to Helicone/helicone development by creating an account on GitHub.

github.com

0

13

Helicone

@helicone_ai

11 months

We are now working with @CommandBar to provide AI assistance and docs directly within Helicone 🎉

1

2

13

Helicone

@helicone_ai

1 month

✌️ Right now, Helicone is the first observability tool with full image support! We currently support: ⃗→ GPT 4 Vision → Claude 3 Opus → Claude 3 Sonnet → Claude 3 Haiku We’re getting more models in the pipeline like Dalle soon - please stay tuned! 👋

3

0

13

Helicone

@helicone_ai

6 months

It finally happened. We "wasted" time implementing dark mode 🌚 🌝 ! The engineering team is excited to release: - dark mode - Helicone Gateway (support for any URL/model) - support for images in OpenAI's gpt-4-vision-preview

2

0

12

Helicone

@helicone_ai

1 year

Bring your ChatGPT API completions to life by visualizing them with Helicone! Wherever you're using OpenAI, get integrated with one line of code

2

1

12

Helicone

@helicone_ai

5 months

Wish you could serve Helicone's dashboard to your users? Now you can! With our Customer Portal, you can... - Serve Helicone directly to your users - Manage user spend - Share custom-built dashboards Please email us at engineering @helicone .ai or DM us :)

0

1

11

Helicone

@helicone_ai

10 months

If you are using Helicone, share your project below and we will retweet (re-x???) it! We will also add your project to our public README.

4

3

12

Helicone

@helicone_ai

6 months

If you are in San Francisco on Dec 2, please come join us at our first Hackathon!

1

5

12

Helicone

@helicone_ai

2 months

Secure your application with 1 line of code. It's as simple as that.

2

1

11

Helicone

@helicone_ai

6 months

🎉 Helicone Gateway + Auto Mapper 🎉 Simply add ANY LLM call through our proxy and we will automatically map your call dynamically to show in our UI. This also enables all the other features you love in Helicone like - Caching - Rate limiting - Dashboards ...etc

0

11

Helicone

@helicone_ai

7 months

We now support @anyscalecompute endpoints

Adithyan

@adithyan_ai

7 months

@Scobleizer @helicone_ai I absolutely love @helicone_ai and the team! And I use that for all my cost measurements. I have also requested them to add the @anyscalecompute endpoints to their monitoring platform.

0

2

11

Helicone

@helicone_ai

9 months

Any other AI companies wanna pivot into super-conductors with us?

0

2

11

Helicone

@helicone_ai

11 months

Helicone dashboard now go brr

1

0

11

Helicone

@helicone_ai

8 months

YC was an incredible and transformative experience. But it's only the beginning of a long journey. Read our short blog post about life after YC and how you can keep the momentum going.

1

0

10

Helicone

@helicone_ai

1 month

@Cloudflare We are super excited to get to know our batch mates for this cohort: @AutoblocksAI , Beam by @MassLuminosity , @bentomlai , @bohr_io , @CommentCloudHQ , Clonable, @diaflow_io , @duliadotio , @erxesHQ , @exporioHQ , Houdini, @IntelligageAI , @LangbaseInc , @go_milkshake , @Nadrama_com ,

1

0

10

Helicone

@helicone_ai

10 months

Introducing async logging! An easy way to capture logs within Helicone without proxying traffic.

1

2

10

Helicone

@helicone_ai

3 months

Within the last month.... 🎉Helicone's cache has saved over 1.6 years of query waiting and over $30,000 for our customers 🎉 Learn how to enable Helicone's cache today below:

3

1

10

Helicone

@helicone_ai

1 year

@supabase 's Large tier -> 8XL Add-on resulted in a 50%+ speed up for most of our queries

0

2

9

Helicone

@helicone_ai

10 months

We’re super excited to see @helicone_ai on @sequoia 's blog post on the new tech stack for LLMs! Huge thanks to @michelle_fradin and @laurenmhreeder for featuring us!

The New Language Model Stack

How companies are bringing AI applications to life

www.sequoiacap.com

0

2

10

Helicone

@helicone_ai

5 months

Automatically link Helicone with OpenPipe to finetune models to improve inference speed, cost, and accuracy. Sign up for our beta here:

0

5

9

Helicone

@helicone_ai

10 months

Exciting news! We have new pricing plans designed for any type of user. Start with our Free Tier: 100k requests/mo & core insights, perfect for getting started. Pro Tier: only $25/mo. For unlimited requests & premium features. Thanks all for your feedback!

0

4

9

Helicone

@helicone_ai

5 months

Congrats to our hackathon winners!! We will be posting the projects soon, stay tuned

1

9

Helicone

@helicone_ai

7 months

Thank you @Scobleizer for having us on!

Robert Scoble

@Scobleizer

7 months

Getting nerdy today. First AI Developer's Tool on my show: @helicone_ai which is a monitoring tool for your AI systems. One change on this third Unaligned: we are more aggressively using @DescriptApp 's AI video editing tools, having it remove extra words and making our guest,

23

44

265

0

9

Helicone

@helicone_ai

4 months

Thank you @wenquai for the shoutout! 🙌

1

0

9

Helicone

@helicone_ai

1 year

LangChainJS support 🚀 LangChian + Helicone edge caching 🧐

LangChain

@LangChainAI

1 year

💪The typescript LLM ecosystem is coming alive All integrations added in last day: 🤗 LLM ( @huggingface ) 📈 LLMOps ( @helicone_ai ) - @justinstorre 🚤 Vectorstore ( @trychroma ) - @jeffreyhuber 🔧 Tools - @upster 📃Document Loader - @dennisamaz 👇

3

24

129

1

9

Helicone

@helicone_ai

5 months

@edels0n No. No way.

0

1

9

Helicone

@helicone_ai

1 month

No more payload size limitations such as embeddings, b64 images, and large token counts. Helicone just levelled up with S3 support! 🙌

0

2

9

Helicone

@helicone_ai

1 year

davinci

1

0

9

Helicone

@helicone_ai

28 days

We’re excited to announce that we are expanding beyond text at Helicone to support image logging. We currently support Claude Vision, GPT-4 Vision, and DALL·E 3. Text and image LLMs are just the beginning as we evolve into a multi-modal observability platform for LLM

0

9

Helicone

@helicone_ai

9 months

Does your team need help setting up Helicone or want a dedicated support channel? Introducing a new support plan! 🎉

0

9

Helicone

@helicone_ai

2 months

🚀 We just added full support for new Claude models, Gemini, and Groq! Cost calculation just got easier.

0

1

8

Helicone

@helicone_ai

7 months

Exciting news for developers! Our latest Node package offers full support for OpenAI v4, with the added benefit of our enhanced logging features. Easily log your OpenAI requests to Helicone via proxy or async and provide feedback with ease! Available here:

1

8

Helicone

@helicone_ai

1 month

Shoutout to @LevanKvirkvelia for helping us add image support for the new GPT-4-Turbo model! We accept and encourage community PRs! If you’d like to contribute, we have put a guide together:

helicone/CONTRIBUTING_GUIDELINES.md at main · Helicone/helicone

Contribute to Helicone/helicone development by creating an account on GitHub.

github.com

0

8

Helicone

@helicone_ai

9 months

The ability to do anything on Helicone is now removed. Why? It makes no sense.

2

8

Helicone

@helicone_ai

9 months

Thank you to everyone who is supporting our Open Source journey 🙏

2

0

8

Helicone

@helicone_ai

2 months

🎉 We’re excited to introduce Lina, our first designer on the team! 🚀 Lina spent time designing in gaming, automotive and fintech before Helicone. She loves storytelling through design and look forward to exploring the nature in SF! Welcome Lina! #NewTeamMember

1

2

8

Helicone

@helicone_ai

5 months

Introducing Cost Alerts! 🌟 Due to popular demand, we're excited to announce that you can now create alerts to manage your expenses. Set up an alert and spare yourself the drama of having to auction off your favorite superhero capes, like I had to. 🦸😭

1

2

8

Helicone

@helicone_ai

1 year

Honored to be part of this incredible list 🙏

Alex Lee ⛷

@AlexLee611

1 year

Here it is. The YC W23 Gen AI Landscape 💥 Why no one had done this yet is beyond me. 36 Gen AI startups in the YC W23 batch. Covering just about everything a company needs: customer success, marketing, developer tools and of course my personal favorite, finance & accounting 🥸

35

153

796

1

0

8

Helicone

@helicone_ai

9 months

This is why we use @discord

2

0

7

Helicone

@helicone_ai

1 year

🎉Helicone just logged its 100,000th request! 🎉 We're thrilled to support the booming LLM community and all the amazing projects they bring to life

2

1

7

Helicone

@helicone_ai

1 year

embeddings

1

0

7

Helicone

@helicone_ai

2 months

JP is one of the most talented engineers we know - we can’t wait to see @hyperaide succeed!

JP

@dqnamo

2 months

Finally using @helicone_ai in a production app! Currently implementing it on @hyperaide and absolutely loving the experience. It's definitely making my life a lot easier.

0

7

2

1

7

Helicone

@helicone_ai

27 days

How it started vs. how it's going (🤫We may or may not have updated our landing page)

0

7

Helicone

@helicone_ai

1 year

We now have a UTC toggle. only took like 1k+ lines of code 😭 😭 😭

2

0

7

Helicone

@helicone_ai

1 year

Unlock insights into your requests and metrics with a low latency proxy between you and OpenAI. Get started quickly and easily, with just one line of code.

1

2

6

Helicone

@helicone_ai

6 months

Helicone is hiring! If you are interested, please dm us or apply using the link below

Helicone: Job Board | Notion

Short link: helicone.ai/career

helicone.notion.site

0

1

7

Helicone

@helicone_ai

8 months

Introducing the new and improved docker-compose. Easily get Helicone running within your infrastructure in minutes

0

1

7

Helicone

@helicone_ai

29 days

Anthropic vs OpenAI time to first token (Accounting for input size)

3

0

7

Helicone

@helicone_ai

9 months

This is the best job in the world

2

0

7

Helicone

@helicone_ai

9 months

are you still not caching your openai api requests? just use helicone:

0

6

Helicone

@helicone_ai

9 months

We want to thank @weekendfund , @vedikaja_in , @rrhoover and @TheManMikeTan for the feature in their Enterprise AI market map! If you're building AI into your products or business, use Helicone for all your observability, monitoring and evaluation needs!

2

0

6

Helicone

@helicone_ai

1 year

Docs:

Introduction - Helicone

How developers build AI applications. Explore our docs to learn how to easily integrate Helicone into your application and explore features that will accelerate your team's development.

docs.helicone.ai

0

1

6

Helicone

@helicone_ai

2 months

👨‍🍳 What’s Helicone cooking this week? 👨‍🍳 We're glad you asked! We added more dashboard graphs due to popular demand. We now have quantiles, time to first token graphs, prompt security dashboard, and threat aggregations overtime, and more. Don’t forget, you can request a new

0

1

6

Helicone

@helicone_ai

22 days

*sees LLM developers who use Helicone for observability and monitoring Us: look between Q and E on our keyboard 👀

1

0

6

Helicone

@helicone_ai

9 months

⚡️Caching in one line of code!⚡️ Caching, by temporarily storing requests, significantly speeds up access time and enhances application performance. 💸Average Savings/User: $210/week 🙅‍♂️Prevent redundant OpenAI requests! 💡Optimize common queries!

0

1

6

Helicone

@helicone_ai

6 months

With @OpenAI 's new models, context windows, and features, its harder than ever to pick whats best for your specific LLM-powered apps. That's why we made updates to our playground. Easily load in your requests and compare different model outputs!

0

2

5

Helicone

@helicone_ai

5 months

🚀🚀🚀 CUSTOM FEEDBACK 🚀🚀🚀 You can add labels/properties to your LLM calls programmatically AFTER the call is made! This lets you easily customize feedback for all your LLM requests! Here are the docs:

Custom Properties - Helicone

Helicone allows you to tag requests with any information you choose to segment, analyze, and visualize by.

docs.helicone.ai

0

6

Helicone

@helicone_ai

6 months

🤩 Update: Helicone Package v2.1.0 The Helicone node package is now at v2.1.0, supporting OpenAI 4.17.0. This includes access to the new beta features. Note: Assistant logging support isn't included yet.

npm: @helicone/helicone

A Node.js wrapper for the OpenAI API that logs all requests to Helicone.. Latest version: 2.1.11, last published: 3 months ago. Start using @helicone/helicone in your project by running `npm i...

www.npmjs.com

2

1

6

Helicone

@helicone_ai

10 months

@sauhaarda Ahh good catch! That’s right

1

0

6

Helicone

@helicone_ai

1 year

We just shipped tagging custom properties to your requests like session, conversation, or prompt_chain id Easily answer questions like: - What is the total latency of a prompt chain of API calls? - What is the average cost of a conversation?

0

1

6

Helicone

@helicone_ai

1 year

@HassanTahir__ Thanks for the suggestion! Done 🚀

0

1

6

Helicone

@helicone_ai

8 months

🚀Helicone Runs🚀 A new way to keep track of your LLM tasks and runs.

2

0

6

Helicone

@helicone_ai

10 months

Come join us at next week’s ClickHouse meet up!

ClickHouse

@ClickHouseDB

10 months

At the upcoming ClickHouse + AI meetup in SF: “Helicone simplifies LLM monitoring, while Clickhouse, integral to our open-source stack, scales effortlessly to handle millions of daily requests.” - @helicone_ai

0

5

16

0

6

Helicone

@helicone_ai

6 months

It’s happening!! Tomorrow!!! Come through 💪

0

6