Introducing AirOps Studio 🌟
Visually build LLM powered workflows, tools, and chat assistants, fast.
We're excited to launch the first version of AirOps Studio! Combine large language models like GPT-4 and Claude 2.0 with your data and documents, transcription services and third party APIs to create solutions that move real numbers in your business.
We've spent the past 6 months speaking to hundreds of companies about their priorities and roadmaps for 2023 and unsurprisingly, LLMs are a huge, recurring theme. The opportunities for LLMs for both user facing and internal use cases are varied and exciting, and unlike traditional software development call for a unique approach. With the launch of AirOps, we are building a platform to help everyone at work create real solutions with LLMs.
We believe that those with the most business context can often be the best prompt writers, regardless of their technical capabilities. We also believe that unlike traditional engineering, building LLM powered solutions is more akin to 'teaching' than 'coding' and therefore requires extensive iteration from prototype through to production. This is why we are building AirOps to unlock your whole team to create and make continuous improvements to AI solutions in simply a matter a clicks.
We soft launched our platform a few weeks ago and our community has already begun creating some amazing things, including:
- Content generation workflows that leverage live web data and proprietary company data.
- Transcription workflows that transform podcasts, meetings, and calls into various written formats, like documents, summaries, and landing pages.
- AI Co-pilots for web products, powered by our Frontend SDK, similar to 'Notion AI'.
- Chat assistants with tools that can be connected to your data and systems.
- APIs capable of generating text, cleaning data, classifying information, and answering queries.
Check out an example here of a Chat Assistant that can compose tweets based on the latest news! It has Tools (built in AirOps) that can search Google, retrieve web pages and then construct a tweet thread based on what it finds. Just give it a topic to create a Tweet thread on!
What's in our v1.0
The AirOps Studio initial release is designed to get you building, today. Here's a rundown:
1. Visual prototyping and building: Combine an expanding selection of 'Steps' to create robust solutions that fuse LLM calls, data, code, and logic. This isn't just another "chat to your PDFs" tool – it's a comprehensive solution for AI development.
2. Managed 'AI Memory' Vector DB: With our managed 'AI Memory', you can easily incorporate your data and documents into your AI solutions. We handle everything from the embedding pipeline to managing the underlying Pinecone DB. We support uploads, Google Docs and SQL today with more coming.
3. Deployment flexibility: AirOps Studio offers multiple ways to use and deploy your applications, including our Frontend Web SDK, Google Sheets Plugin, and batch runner via CSV. We want to make it easy to bring our solutions to surfaces that matter most to you.
4. Versioning, monitoring, and scaling: Given that LLM deployments aren't a one-and done, we've incorporated features like app versioning (with a versioned API), full execution logs, and bulk evaluation to make it simple for you to manage and scale your solutions.
This launch comes off the back of our recent $7m fundraise. We're putting this to work building the best place for you to create solutions with LLMs.
Have questions? Don't hesitate to email us at email@example.com or chat with us in-app.