News Cache (2026-03-02)

February 21, 2026 - February 28, 2026


Top Highlights

  • Vercel Queues enters public beta as a durable event streaming system built with Fluid compute, enabling direct message publishing and consumption alongside Workflow for multi-step orchestration.
  • New Chat SDK launches with a unified codebase approach for building chatbots across multiple platforms including Slack, Discord, GitHub, Teams, and now Telegram, eliminating the need for separate implementations per platform.
  • Vercel Sandbox introduces HTTP header injection for secure credential management, allowing sandboxed code to call authenticated services without ever accessing API keys or tokens directly.
  • Security boundaries in agentic architectures explores the critical vulnerability of most agents today running generated code with full access to secrets, providing technical insights for building more secure AI systems.

Community Projects

Don’t forget to visit the Showcase to see what else everyone is building

Weekly Leaderboard

Congrats to last week’s leaderboard champions

  1. @tulioportela
  2. @abhivarde
  3. @marnix-3108

Resources

Events

More Updates

5 Likes

Grateful to be part of this community :raising_hands:
I’ve been shipping consistently for months and building in public on Vercel

Looking forward to continuing the work and seeing where it leads :eyes:

2 Likes

Nice ship again! Cant hate on security releases. Good work​:flexed_biceps:

1 Like

Thanks for featuring AI Video Generation Playground Project!

1 Like

Can’t wait to start using queue - but it does not seem to work for nuxt. Following instructions on how to set up vercel.json breaks build and deploy is impossible when setting up vercel.json as defined in documentation.

2 Likes

Saw your post, @ojpetar! I let the team know :slight_smile:

1 Like

Thank you for your announcement.
May I ask some question?

How does Vercel Queues differ from traditional queue systems like Apache Kafka, RabbitMQ, or Amazon SQS?
What are the durability guarantees of Vercel Queues?
Can Vercel Queues support high-volume event streaming workloads, or is it mainly designed for background jobs?
How does Fluid compute improve performance or scalability for event processing?
ow does Workflow orchestration integrate with Vercel Queues in multi-step pipelines?