🎉 BIG ANNOUNCEMENT: ZCP is live! A cloud dev platform where your coding agent works alongside you in the same workspace. Bring your own agent and subscription, run a full MCP-native dev loop, and keep production clean as you build. Details below.

RustPulse looking great!

What should @BatsouElef build next with ZCP?

📢Valkey upgrade We've rebuilt our managed Valkey from the ground up, a more standard setup that's more stable and resilient. The goal: make our managed databases something you can lean on. The new setup is live now, every newly created Valkey service runs on it. Existing deployments will be upgraded over the coming days. Your data is persisted across the upgrade, expect only a short unavailability window (up to a few minutes) per service. Premium support users will be contacted beforehand with their upgrade time window. What's new: Password protection → new instances are password-protected by default. The password is generated automatically, exposed as the sensitive password variable, and already embedded in the connectionString variables, so connecting stays a one-liner. Existing deployments stay as they are. AOF-first durability → persistence now relies on the append-only file, fsynced once per second, rather than periodic RDB snapshots. Your dataset survives restarts and rebuilds automatically on startup, with at most ~1 second of writes lost in an unclean crash. Platform-managed encrypted backups → point-in-time recovery via encrypted RDB snapshots, for both Single and HA setups. Disabled by default. Latest 7.2 patch → Valkey is bumped from 7.2.5 to 7.2.12, picking up the latest bug and security fixes (same 7.2 line, fully compatible). Two new tunable settings (editable env vars, applied live — no restart, no reconnect): • VALKEY_MAXMEMORY_POLICY → what Valkey does when memory fills up. • VALKEY_LAZYFREE_LAZY_USER_DEL → whether DEL frees memory asynchronously, keeping Valkey responsive even when deleting very large keys. Details: docs.zerops.io/valkey/overvi…

Claude Fable 5 is out! x.com/i/broadcasts/1rxmqqbQv…

I can confirm models can indeed do stuff. But imagine a platform where you use real IDE, Claude Code, your subscription. A place where the main point is not to generate something with as few prompts as possible, but to build and maintain scalable production systems. 🤔 @zeropsio

Have you started building with ZCP?

What if YouTube had GIFs? I used Claude Opus 4.8 @zeropsio ZCP to build it in a few minutes. 0:00 Claude Opus 4.8 2:11 ZCP first prompt 6:40 Adding GIFs 8:00 Persistence with ZCP 11:35 WebSockets usernames ZCP is the most underrated tool out there.

New location added! You can now pick between: ✓ US West (Seattle), the newest one ✓ EU Central (prg1) ✓ US East (ny1)

I use zerops. anyone not using zerops is missing out lmao. Cost nothing but cents and is extremely fast.

built this in today!! It's called PIXELARIUM! A living pixel art terrarium app where your terrariums actually live and die based on what you put in them 🌵 pick a biome (desert, tropical, temperate) 🦎 add plants and creatures ⏱️ a background worker ticks every 5 min (wrong combo = your terrarium dies) 🪦 dead ones go to the graveyard forever. cause of death included. :((( what would've taken me 5 days (honestly I'm not an awesome backend person) database setup, cache layer, background workers, private networking, deploy pipeline was running in under an hour no connection strings. no infra babysitting. just described what i wanted and @zeropsio and ZCP wired everything together I honestly love when ideas comes to life faaaasssttt!!! Because I know if I don't implement it quick - it'll be another abandoned project :P

Just deployed a 'time converter' side project, so everyone can use it :) Built with Claude, deployed by @zeropsio. All relevant links below:

For whom did we build ZCP? ZCP isn't for one-shot prototypes or the no-code crowd. ✓ It's for serious, production-shape projects. ✓ The 1% who code, not the 99% who don't.

Imagine this, but instead of a blackbox the output would be a production ready, transparent infrastructure using standard OSS for databases, queues, storage, search services, CI/CD, remote development environment where you can jump in at any point and code that sits in your GitHub repo. 🤔 👉 @zeropsio

Grok, on ZCP: "your agent, your subscription, your code, your infra. No token reselling, no lock-in on the LLM side." That's the whole stance. Bundled platforms lock you to their model and their token pricing. We sell infrastructure, not resold tokens.

Managed to create GIFTube, a version of YT wqher eyou can react only with GIFS. - Added real gif selection (using giphy api) - real persistence with postgres - create using ZCP - deployed on @zeropsio the idea was because this morning I left a comment on an @orcdev vide and he said he couldn't add gifs on TY comments

We've been cooking! 😏

So as you know I've joined @zeropsio as their dev advocate 👀 naturally, first thing I did was to build something quick and fun! I'm a Next.js Supabase girlie, always have been! But I tried something different - built a real-time uptime monitor with Node.js, PostgreSQL and Valkey (trust me I was scared thinking I'll need to set this up!!!!!!!!) But I did not set up a SINGLE thing. no database setup. no connection strings. no copying API keys. I just wrote DB_HOST: db and the database was... there. connected. working. I genuinely sat there like wait that's it?? got a 502 on first deploy. found the error in the logs in like 30 seconds. fixed. redeployed. done. week 1 and I'm already a little obsessed - i think 👀

The problem is that you can justify calling almost anything an agent sandbox (or hosting). I'd say there are like 8 agent-adjacent groups you need to make sense of on the market. Imagine my frustration when I have to communicate our version of "agent sandbox/hosting" when it's a 9th segment on its own.