531 posts tagged with "AI agents"
AI agents and autonomous systems
bex Is in Public Alpha: What Ships Today, What It Costs, and What 'Alpha' Means for the Open-Source, Agent-Operable Render Alternative
bex is open for signup as a public alpha: git push to HTTPS on machines you own, a Render-compatible API and CLI, 179 MCP tools for agents, and prices 30% under Render. Here is the ledger of what ships, what is missing, what it costs, and what must be true before 1.0.
Your AI Agent Can Now Restart Prod Postgres From Chat: What the Coolify–AnythingLLM MCP Bridge Costs in Audit and Rollback
An open-source MCP server lets an AI agent deploy, restart, and roll back Coolify apps from an AnythingLLM chat window. A threat-model teardown of what agent-driven deploys cost in scope, attribution, and reversibility — and the five-item checklist any agent-facing deploy server must pass.
Crossplane and AI: Why Agent-Operated Platforms Need Declarative APIs, Not Dashboards
CNCF's case that AI agents need API-first declarative infrastructure is right — but Crossplane's sprawling CRD surface is the wrong window into it. Agents need a small typed PaaS API over declarative machinery.
Dokploy Shipped 508 MCP Tools. The Community Deleted 95% of Them.
Dokploy's official MCP server mirrors every API endpoint as a tool — 508 of them, burning roughly 74k tokens before the first question. Community rivals cover the same API with 13 to 28 curated tools. The measured numbers behind the backlash, and five design rules for the next infrastructure MCP.
Building Agents That Don't Break Themselves: Fly.io's Sprite Safety Patterns and the Deploy-from-Chat Risk Model
Fly.io's June 2026 Sprite safety patterns — splitting the agent loop from execution, per-run credentials, and checkpoint restores — map directly onto what a deploy platform needs before agents can ship code: scoped credentials, audit logs, and guaranteed rollback.
Kubernetes Gets Its Own Agent Sandbox: What the Sandbox CRD Means for a Self-Hosted PaaS
Kubernetes SIG Apps shipped a Sandbox CRD for running AI agents as isolated, stateful workloads. Here is what warm pools, snapshot resume, and pluggable gVisor/Kata isolation mean for teams running their own PaaS, and the gaps to track before adopting it.
Pinterest's MCP Servers Save 7,000 Engineering Hours a Month: What 66,000 Invocations Teach About Agent-Operated Infrastructure ROI
Pinterest's production MCP ecosystem handles 66,000 monthly tool calls across 844 engineers, saving an estimated 7,000 hours a month. A teardown of where those savings came from, how solid the math is, and which operations a smaller platform should expose to agents first.
Pion Wants to Run Your Company Fully Autonomously: What Andon Labs' September 2026 Launch Assumes About the Infrastructure Underneath an Agent With a Bank Account
Andon Labs' Pion hands persistent AI agents a bank account, a terminal, and full control of real businesses. The Vending-Bench slope, the $38k cafe loss, and the infrastructure checklist — credentials, durable state, budgets, audit — a platform must build before agents get production keys.
Render's MCP Server Grows Teeth: OAuth Plus trigger_deploy Set the Agent-Deploy Bar
Render's July 2026 changelog added OAuth login and a trigger_deploy tool to its MCP server. This breakdown tables the full agent-deploy surface, names where Render's own docs stop, and lists what a self-hosted PaaS must ship to match it.