About

OpenSay is a Slack app designed to help integrate anonymity and whistleblowing within organizations in a responsible manner. It is trusted worldwide by more than 2500 teams and utilizes AI language analysis alongside team oversight for moderating content. The app supports anonymous communication in both public channels like Ask Me Anything and Suggestion Box, and private channels including Whisper to HR and Whistleblower Hotline.

OpenSay’s mission to facilitate discussions on contrarian viewpoints and delicate issues seeks to elevate transparency in the workplace, broaden the spectrum of thought, and cultivate a culture that is both psychologically safe and healthier.

How they use Cloudflare

OpenSay leverages Cloudflare’s scalable serverless architecture to build a monolith service on the developer platform. To reduce overhead and time spent debugging microservice issues, OpenSay compresses numerous libraries to a single master Worker to offer a layer of observability via distributed tracing into the OpenSay architecture. The master Worker then compiles all the microservices to run the main logic of the application.

OpenSay uses Cloudflare Edge network to improve the response times of their app resulting in better UX (see blog post).

To ensure images propagated globally, OpenSay built a cache feature that runs on Cloudflare Cache When requests for an image is sent to a Worker, the Worker then requests an image from a microservice, returning the binds of the image, then adding response headers to propagate the cache.

OpenSay also uses the HuggingFace integration with Workers AI to test language models like Llama Guard for specific user-based behaviors.

Why Cloudflare

“If you have not worked with a lot of GCP / Lambda functions, a developer wouldn't know the pain or costs associated. Workers make serverless deployments just so much easier.”

“Before Cloudflare Workers, I didn't want to use AI due to costs, but now that Workers AI is in the Cloudflare ecosystem, it's so easy to run a model.”