Parteek Singh

docs can be for humans, too :)

With Agentuity sandboxes, the same prompt can generate TypeScript and Python code, run both in isolated runtimes, and use background evals to verify outputs. @ParteekS_ walks through a quick example to show you how it all works. See the full demo: youtu.be/sfxaG9jmbF4 Check out the code: github.com/agentuity/exam…

New database performance tab in the dashboard, from @jhaynie. Click on any database and get intelligent query insights, slow query detection, latency and volume trends, and more. Check it out: app.agentuity.com

Most agents need more than a chat interface. They run on schedules, handle webhooks, and integrate with external services. On Agentuity, every trigger is a route. Your agent logic stays separate from how it gets called, so you can add integrations without changing your agent code. Get started today: app.agentuity.com

Agents need APIs, and routes on Agentuity are standard @honojs. Define your endpoints, add middleware, and deploy alongside your agents. Every route handler gets the same cloud services your agents use. Your agent schema doubles as route validation, and the frontend gets a typed hook that stays in sync. Learn more: agentuity.dev/routes/http Get started: app.agentuity.com

We built a Slack and Discord bot from a single Agentuity agent using Vercel's Chat SDK. Conversation history lives in KV with a sliding window, and our AI Gateway handles the model calls. Check out the code: github.com/agentuity/exam…

Football mock drafts can sometimes suck. @agentuity agents can make them fun 🤖 I used our new @opencode and Claude Code plugins to create a draft simulator with a team of agents. See the full demo: youtu.be/yJaRJKEQafM Check out the code: github.com/parteeksingh24…

We just open sourced a full-stack web app that spins up sandboxes for coding agents, all built on the Agentuity platform 🤖 Each session gets its own sandbox running @opencode. Save snapshots, fork sessions, SSH in, or attach locally. Everything syncs in real time through durable streams. Ships with Agentuity Coder: a full agent team backed by our cloud memory, with built-in Cadence mode for long-running tasks. Try it out: github.com/agentuity/code…

We built an entire dev team of agents to help us build our product. Now it's available as both a Claude Code and @opencode plugin. Agentuity Coder extends our CLI with agents backed by cloud memory (KV + vector store). They remember what worked, what broke, and what had to be fixed. Works across sessions and across your team. See it in action: youtu.be/m6n6mb98kgs Learn more: agentuity.dev/Reference/CLI/…

v1 is here, and we're ready to show it off! HackAI brings together builders in Austin to learn and build with AI. We'll be there tomorrow doing a demo. Come say hello! Details: meetup.com/hack-ai/events…

Your agents run in production, and they do more than just make LLM calls. Your evals should too. Agentuity evals cover every session, across the entire run, automatically.

Loxahatchee River District treats wastewater and recycles it for irrigation, preserving fresh water for the environment. Now they're deploying agents to work alongside staff. "I can have an employee, which is really an agent, off-site, responding back with solutions, no different than a junior staff member." — Albrey Arrington, Executive Director

@alexcloudstar Great question! We think complexity is inherent when building with agents, so we handle scaling vertically (more resources per agent) and horizontally (more agents) with primitives like queues and sandboxes.

@neondatabase @agentuity Sounds promising! Curious how it handles scaling with complex workflows 🤔

Existing cloud infrastructure was built before agents were doing most of the work. If we want to unblock agents, they need a cloud made for them. That’s what @agentuity is launching: a full-stack platform with primitives designed around how agents actually run and manage state

Agentuity v1 is live! 🚀 A cloud purpose-built for AI agents. Full-stack development, built-in services, sandboxes, production evals, and more (🧵)

Join us for the v1 Launch video premiere on Monday at 9am PST (5pm UTC). We're excited to share all we've been building to make Agentuity the first and only full-stack platform built for AI agents. youtube.com/watch?v=-IklHv…

The new native sandbox support built into the @agentuity v1 platform. Showing off a really cool new OSS project called agent-browser.dev running inside. The future is agentic.

Agentuity v1 has reached beta! 🚀 We've been shipping fast since the public preview in December: sandboxes, agent evals, built-in auth, and more. (more in 🧵)

End-to-end type safety in v1 🚀 Define your schema once, and types flow through your agent, API route, and frontend automatically. Give it a spin: curl -sSL v1.agentuity.sh | sh #AI #AIAgents #TypeScript #DevTools

With the Agentuity v1 platform, you can build multi-turn chat agents with built-in memory. Use ctx.thread.state to persist conversation history across requests. No external database needed. Install the CLI to get started: curl -sSL v1.agentuity.sh | sh (more details below)

Real-time LLM streaming with SSE in v1! It takes just a few lines of code: define an endpoint with `router.sse()`, pipe tokens (here using the AI SDK), and stream them to your client. Try it out today: curl -sSL v1.agentuity.sh | sh (more below 👇)