{"id":95573,"title":"Castari - Vercel for AI Agents","tagline":"Deploy your sandboxed agents with Castari","body":"Hey! We’re Jacob Wright \u0026 Cambree Bernkopf, founders of [Castari](https://castari.com).\n\n### **TL;DR**\n\n**Castari lets teams deploy AI agents to secure, autoscaling sandboxes in seconds.**\n\nYou write agent logic with the **Claude Agent SDK (other frameworks coming soon)**. We turn it into a production-grade endpoint.\n\nWe’re especially focused on:\n\n* Teams building **internal agents** (ops, support, data, eng workflows).\n* Companies whose **core product is an agent** built on Claude Agent SDK and need a safe, reliable way to run it.\n\n### Launch Video:\n\nhttps://youtu.be/81K80OivtAc\n\n### The Problem:\n\nSerious agents are not just a single API call. [Agents need a computer.](https://www.anthropic.com/engineering/building-agents-with-the-claude-agent-sdk)\n\nThey execute code, call tools and browsers, run for minutes instead of milliseconds, and touch sensitive systems. Most teams end up rebuilding the same fragile setup:\n\n* Homegrown agent harnesses\n* Sandboxes, file systems, and queues stitched together\n* Logs and traces scattered across many services\n* Ad hoc snapshots and risky secret handling\n\nThe result:\n\n* Every agent needs its own mini runtime\n* Debugging one agent run feels like detective work\n* Adding a new agent feels like starting a new infra project\n\nGood agents stall at “cool demo” instead of reaching production. Deploying agents becomes a full-time infrastructure problem.\n\n![uploaded image](/media/?type=post\u0026id=95573\u0026key=user_uploads/89898/44917ba7-819c-4c41-a80f-ef359054b19e)\n\n### How Castari solves it\n\nWe’re building Castari to be the **natural runtime for AI agents built on Claude Agent SDK**. We wrap the entire agent in a sandbox, deploy it, and provide you with an endpoint url that auto-scales as requests come in. No need to write code in your tools to manage your sandbox lifecycle.\n\n#### **1. Drop in a config file → get a production runtime**\n\nIn your Claude Agent SDK repo:\n\n* Add a `castari.yaml` (entrypoint, tools, env vars)\n* Run `castari deploy`\n\nThat’s it. Your agent is now running in a secure sandbox with an endpoint, UI, and observability.\n\nWhether that agent is:\n\n* an **internal system** your team uses\n* a **customer-facing agent** that _is_ your product\n\nthe deployment workflow is identical and boring (in the good way).\n\n#### **2. Sandboxing \u0026 safety, built-in**\n\n* Every run executes in an isolated sandbox designed for tool-using agents\n* You design the behavior; we handle isolation and scaling\n* No more duct-taping together sandbox primitives for specific tools\n\n**Bottom line:** You keep building on **Claude Agent SDK** the way you already are — for internal workflows or core products. Castari turns that into a safe, observable, production-ready runtime.\n\n### Why Castari is different\n\n**E2B, Modal, Daytona, Cloudflare** (and friends) give you powerful sandbox and compute primitives.\n\n**Castari sits one layer up:**\n\nCastari wraps around your Claude Agent SDK-based agents so you can define them declaratively and run them in secure sandboxes, without owning the underlying sandbox lifecycle.\n\n* **Agent-first semantics**: runs, sessions, tools, snapshots are native concepts we manage.\n* **Zero framework lock-in**: remove Castari and your agent still runs on vanilla Claude Agent SDK; keep Castari and your infra team gets its life back.\n* **Focused start**: deeply integrated with **Claude Agent SDK** today, with the same “drop-in config” experience coming to other frameworks next.\n\nWe’re building the runtime we wish existed for agents.\n\n### Why we’re working on this\n\nWhile I was leading the FDE team at RunPod:\n\n* Built and shipped AI infrastructure and agents for **Fortune 500s and fast-growing startups**\n* Watched the same loop repeat: agent works in a notebook → weeks disappear into infra glue → nobody trusts the system, production release gets delayed (over and over again).\n\nAdditionally, while [**@Cambree Bernkopf**](https://bookface.ycombinator.com/user/1372064) and I scaled our AI consumer app to 2m+ members, our biggest issues were getting our agents to work reliably in production.\n\nThe pattern is clear: **agents deserve a first-class runtime, the way frontends got one.**\n\nCastari is our bet that:\n\n* Sandboxing \u0026 safety should be built-in, not bolted on\n* Going from “prototype” to “production-safe” should be one config file, not a quarter-long project\n* AI teams should ship agents, not infrastructure.\n\n![uploaded image](/media/?type=post\u0026id=95573\u0026key=user_uploads/1070542/5fbe422a-1cc4-49e3-95ce-8740164b600e)\n\n### Ask: early pilot access\n\nWe are opening a limited pilot for:\n\n* Teams building internal agents who need a safer, more reliable runtime\n* Companies building agent centric products that need to run in production without bespoke infra\n\nIf that is you and you are tired of wrestling with sandboxes, observability, and brittle tooling:\n\n→ Email [**jacob@castari.com**](mailto:jacob@castari.com) with subject “Castari Pilot” and 2 to 3 sentences about your agent stack and use case.\\\n→ Check out our open source repo for using any model with Claude Agent SDK on [GitHub](https://github.com/castar-ventures/castari-proxy).\\\n→ Join the waitlist on [**castari.com**](http://castari.com) to hear about our first releases.\n\nJacob, Cambree, and the Castari team","slug":"OrV-castari-vercel-for-ai-agents","created_at":"2025-11-17T20:28:20.453Z","updated_at":"2026-05-25T01:54:21.356Z","total_vote_count":20,"url":"https://www.ycombinator.com/launches/OrV-castari-vercel-for-ai-agents","share_image_url":"https://www.ycombinator.com/media/?type=post\u0026id=95573\u0026key=user_uploads/89898/44917ba7-819c-4c41-a80f-ef359054b19e","company":{"id":30975,"name":"MadeThis","slug":"madethis","url":"https://madethis.com","logo":"https://bookface-images.s3.amazonaws.com/small_logos/73b59fa186ca146c4f90522806e84fb89b3f9f10.png","batch":"Fall 2025","industry":"B2B","tags":["Artificial Intelligence","Operations"],"search_path":"https://bookface.ycombinator.com/company/30975"}}