Big announcement coming next week ๐
Beam
138 posts
Open-source cloud platform for AI inference, sandboxes, and agents (YC W22)
- It's Day 3 in launch week and we're shipping our biggest feature of the week Today we're releasing Beam Bots ๐ค It's a new bot framework with sandboxed compute and concurrency built-in. Shoutout to all our friends participating in launch week with us @supabase, @PropelAuth,
- ๐ Launch Week, Day 2๏ธโฃ It's Day 2 in our MEGA LAUNCH WEEK with @supabase @PropelAuth @magicpatterns @JamsocketHQ and more. Today we're releasing realtime apps โจ
- One feature. Every day. For five days. We're joining the @supabase MEGA LAUNCH WEEK next week from December 2nd-6th. We've got some crazy stuff planned, things like... ๐ฅ Massively parallel armies of AI agents ๐ฅ Ultra low-latency inference for real-time apps ๐ฅ Managing
- Attention Python enthusiasts in SF ๐ We're co-hosting a meetup with @getreflex this Thursday 6/12 Learn how to build full-stack AI apps entirely in Python, and get a behind-the-scenes look at the infrastructure that makes it possible Spots are limited, so make sure to RSVP
- โจ Launch Week, Day 5๏ธโฃ It's the final day of MEGA LAUNCH WEEK with @supabase, @PropelAuth, @magicpatterns, @JamsocketHQ, @triggerdotdev, @pinecone and many more. Today we're shipping a CLI for our self-hosted users to manage their Beam cluster. ๐ Details below
- Have you tried running our ComfyUI workflow? ๐ค $30 free credit when you signup ๐ GPUs from 4090s to H100s ๐๏ธ Mount storage volumes to save files ๐ Ship your Comfy workflow as an API Give it a try! beam.cloud
00:00 - โจIntroducing the Official Beam MCP server Instantly give Claude and Cursor full access to Beamโs cloud platform with a single text prompt. Launch scalable GPU apps, spin up code sandboxes, or deploy a complete AI backendโit's as easy as texting a friend!
00:00 - ๐ Launch Week, Day 1๏ธโฃ Today we're releasing our new vLLM engine. We've built @vllm_project directly into the Beam SDK: โ Run vLLM models on the cloud with one command โ Deploy vLLM inference endpoints โ Scale out to 100s of GPUs โ Automatic scale-to-zero
- We just shipped a new autoscaler for managing high-traffic GPU workloads. If you want to scale based on the number of tasks in your queue, just add this autoscaler to any function on Beam. ๐ Learn more about autoscaling here: docs.beam.cloud/deployment/autโฆ
- โจ New Feaure: Warmup Containers We shipped a new API to warmup your containers. This API will: - spin up a container - load your app + model weights into memory - warm up the container without launching a task There's also a new "warm up" button on the dashboard. Give it a
- New: Separate Environments with Workspaces โจ Workspaces allow you to create individual environments in Beam. For example, you can create separate workspaces for development, staging, and production. It's live now! Login to your dashboard and give it a try ๐
GIF - Build an AI-powered Youtube Transcription app with @supabase @deno_land and @beam_cloud by following this tutorial! ๐ฅ







