radhakrsna

Hi HN,

Today, we are excited to be open-sourcing GPTRouter, an LLMOps tool we have been using internally at Writesonic for handling millions of monthly requests for our users.

Universal API for 30+ LLMs, Vision and Image Models Smart Fallbacks based on latency and uptime Automatic Retries Supports streaming

Since embracing OpenAI GPT-3 in production in 2020, we at Writesonic have been serving millions of users and faced the typical scaling pains with generative AI models:

1. Dependency on a single model risked total downtime. 2. Latency issues with models like GPT-4 affected user experience. 3. Integrating various models was tough due to different APIs and SDKs.

Early this year at Writesonic, we set out with a clear vision: to become model agnostic.

Faced with single-model limitations and diverse AI challenges, we began building GPTRouter - our bespoke solution to navigate and thrive in a multi-model AI world.

With GPTRouter's Universal API, you're the master of AI models. Swap between OpenAI, Azure, Anthropic, Replicate, Cohere & more with just one line of code. It simplifies model management to a great extent.

Downtime isn't an option. GPTRouter's Smart Fallbacks mean your service is always on. You can define a hierarchy of models for each use case. GPTRouter will constantly check for uptime/downtime, latency and other factors, and automatically fallback to the next best model with zero interruption.

Say goodbye to manual retries. GPTRouter does the heavy lifting with Automatic Retries for failed requests, keeping your AI services sharp and consistent.

GPTRouter's Edge: Universal API for seamless model switching. Smart, automatic fallbacks for continuous service. Reduced latencies for quick interactions.

Additionally, we will also be open sourcing our frontend LLMOps layer that provides a playground to test multiple models in parallel, keep a tab on the latencies for each model, track tokens and costs for each model and user all in one place.

We are looking forward to seeing how developers leverage GPTRouter in their own use cases.

Thank you!

Yup, do try it out. It's Writesonic lyrics generator.

For sure :)

Glad you liked it!

We used artificial intelligence to generate an entire music video in the style of Eminem.

First, we used Writesonic which is an AI-powered writing assistant to generate the lyrics for the song. We used the phrase "I'm gonna change the world" as the seed text and music style as Rap. Writesonic's song lyrics generator was able to generate multiple different variations of lyrics from just this one line of input. We picked the one we liked the best.

Secondly, we used an AI model trained on Eminem's vocals to generate the voiceover for the lyrics.

Then, we used another AI model to generate background music. We synced up the vocals and the music.

Finally, we used OpenAI's Dall-E 2 AI to generate images for each lyric in the song and combined all of them into a video.

- Lyrics: Writesonic's AI Lyrics Generator

- Vocals: AI trained on Eminem's music

- Beats: Another AI Model

- Images: OpenAI's Dall-E

Here are the lyrics:

I am gonna change the world

With my music, my words

I'll make them listen

I'll make them hear

The message that I bring

It's time for a change

And I'm gonna be the one

To lead the way

I'll be the voice for the voiceless

The one who's not afraid to stand up

And be counted

I'll make a difference

In this world

With my music, my words

I'll make them listen

I'll make them hear

The message that I bring

It's time for a change

And I'm gonna be the one

To lead the way

Do you want to join me?

Do you want to make a difference too?

If you feel like it's time for a change

And you're ready to stand up and be counted

Then come with me

And together we'll make a difference

In this world

With our music, our words

We'll make them listen

We'll make them hear

The message that we bring

It's time for a change

And we're gonna be the ones

To lead the way

Try out Writesonic's AI Lyrics Generator: https://writesonic.com/

Hey everyone!

According to a McKinsey analysis, the average professional spends 28% of the workday reading and answering email. Imagine if you could get that time back!

That’s why I built Magic Email

Magic Email is an AI Email assistant that uses GPT-3 to let you:

1. summarize long emails into short, readable summaries

2. generate professional emails from one-line descriptions

3. generate contextual replies from one-liners or key points

4. auto-correct grammatical errors in your email

Join the beta here: https://magicemail.io

We’d love to get your feedback and look forward to answering any questions!

Hi, thank you for appreciating our product and landing page.

There are quite a few articles which are saying that touchless tech will witness growth much faster than earlier due to the COVID-19 outbreak.

Some links: https://www.securitymagazine.com/articles/92823-covid-19-and... https://builtin.com/design-ux/future-touchless

Hey!

Touch-less technology is the future (especially after the Covid-19 pandemic).

So, I made Zesture, a Mac/Windows app that uses your laptop's camera to give you touch-free control over your media, entertainment and presentation applications (without any extra hardware).

You can watch a demo video here - https://www.youtube.com/watch?v=-_swm09Xmtg&feature=emb_titl...

Supported Apps and Websites:

- Music: Spotify, Apple Music, Amazon Music, Spotify (Web Player), YouTube Music, Amazon Music (Web Player) and Deezer.

- Video: VLC Media Player, YouTube and Netflix.

- Presentation: Microsoft PowerPoint and Keynote.

️Supported Actions: Play/Pause, Next/Previous Track, Enter/Exit Full-screen, Forward/Rewind, Mute/Unmute, Volume Up/Down.

Privacy: Using your webcam is totally secure. Gesture recognition is done locally on your computer and hence we don't record, save or send any images or videos at all. Your camera is turned off automatically after a certain period of inactivity (configurable via the app).

We’d love to get your feedback and look forward to answering any questions!

True, there are quite a few similar services but not many seem to work well. Our service provides better summarization (at least for the articles I tested), had additional features like extracting author name, publish data, important keywords etc and also comes with browsers extensions so you could summarize pages at the click of a button.

The method you described is a part of our algorithms but more steps are needed to make it give meaningful results and make sure it works on different kinds of articles.

Yes, you are right. There is still room for improvement and we will keep trying to make it better. The reason I added paid plans was to test whether people would be willing to pay for such a service so that I can spend more time in adding more features to it and making it better.

Thank you! Yes, it doesn't work that well on technical articles. We will try to keep improving it.

Glad that you liked it. Have you tested it out with the advanced summarizer? Thank you very much for your encouragement. I will try to keep improving it.

You could try the advanced summarizer. It gives better results.

Can you please let me know the article that you tested it on? Maybe you could try the advanced summarizer and see if it gives useful results.

It isn't really built for manuals etc. It is mainly for blog posts or news articles.

The Basic Summarizer has its restrictions. Try Advanced summarizer. It will give better results.

Glad you liked it. Let me know if you have any feedback/suggestions. Yes, we do plan to launch an API soon. Please message us here - https://tldrthis.com/contact and we will let you know when we launch it.

v1 of our service was free as well. v2 includes a basic summarizer which is free and an advanced summarizer which requires payment. Just launched the premium version, so waiting to see.

Yes, that's because that sentence appears twice in the article. Advanced summarizer gives a better summary for that URL.

Thank you!