{"id":30682,"date":"2026-05-07T08:28:23","date_gmt":"2026-05-07T08:28:23","guid":{"rendered":"https:\/\/www.europesays.com\/ai\/30682\/"},"modified":"2026-05-07T08:28:23","modified_gmt":"2026-05-07T08:28:23","slug":"why-long-running-ai-agents-break-on-http-and-how-ably-is-fixing-it-the-new-stack","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ai\/30682\/","title":{"rendered":"Why long-running AI agents break on HTTP and how Ably is fixing it – The New Stack"},"content":{"rendered":"

For this episode of The New Stack Makers, we speak with Matthew O\u2019Riordan<\/a>, CEO and co-founder of Ably<\/a>. <\/p>\n

Ably\u2019s platform handles trillions of transactions a month for customers like HubSpot and Intercom. Originally built for human collaboration, its technology now also nicely enables long-running AI agents.<\/p>\n

\u201cWe\u2019re just very well suited to solving this communication layer between agents and humans.\u201d<\/p>\n

In this episode, we discuss why HTTP starts to break for long-running agents, how a \u201cdurable session\u201d layer is emerging in AI infrastructure, and why Ably wants developers to stop thinking about pub\/sub.<\/p>\n

The AI bandwagon Ably tried not to jump on<\/p>\n

About 18 months ago, O\u2019Riordan says, investors and his own team started pressing him for an AI story, but he initially resisted. Ably wasn\u2019t an AI company, after all. It was an infrastructure play built for human collaboration with a focus on presence, ordering, state, and reconnection.<\/p>\n

Then agents came along. Once those became long-running processes that reason and call tools, they also needed to stay alive for hours<\/p>\n

\u201cThis is what we\u2019ve been building for humans to communicate. And agents are becoming more human-like. [\u2026] I\u2019d love to say we had the vision, and that\u2019s why we built it all, but it\u2019s not,\u201d O\u2019Riordan says. \u201cWe\u2019re just very well suited to solving this communication layer between agents and humans.\u201d<\/p>\n

Where HTTP starts to break for agents<\/p>\n

Most AI applications start on HTTP. It\u2019s the default for developers on frameworks like LangGraph, and as even Ably would concede, it\u2019s just fine for short, one-shot completions from chatbots. But a long-running agent that calls dozens of tool calls over multiple reasoning steps creates problems that a standard request\/response flow wasn\u2019t designed for. The system needs to handle connection drops, users switching tabs or devices, or interruptions by users wanting to interrupt the agent mid-stream.<\/p>\n

HTTP isn\u2019t the wrong transport tool, O\u2019Riordan says. \u201cIt is exactly what you need to get up and running.\u201d But expectations have shifted because, as he puts it, \u201cwe\u2019re all engaging with ChatGPT and Claude.\u201d<\/p>\n

\u201cWhat most people are doing is they click between tabs and they just expect the experience,\u201d he says. \u201cYou ask it something, it takes a while, you go away, you come back five minutes later to a different tab, and you expect everything to just be in sync.\u201d<\/p>\n

The vocabulary developers have started using for this layer is \u201cdurable sessions.\u201d O\u2019Riordan credits EMQX<\/a>, the MQTT broker, with originating the term and ElectricSQL<\/a> with popularizing it for AI. He prefers it over \u201cdurable streams\u201d because streams are only part of what\u2019s needed. A durable session also covers presence, shared state, storage to rehydrate after crashes, and push notifications for offline users.<\/p>\n

To support this AI use case, Ably extended its primitives, O\u2019Riordan says. Token streaming requires moving from immutable to mutable messages, so that a client reconnecting mid-stream can request the current state rather than replaying every token. The company also introduced \u201clive objects,\u201d collaborative storage that keeps an agent and a human in sync on shared state, even when the agent has reasoned for so long that the user has moved on.\u00a0<\/p>\n

The drop-in replacement<\/p>\n

The other constraint O\u2019Riordan kept hearing from developers was that any new transport layer had to be effectively invisible. The goal, he says, was a layer developers could drop into their existing infrastructure without changing their overall stack and design patterns.<\/p>\n

Ably\u2019s answer is AI Transport<\/a>, which is now in its second iteration. It leaves the client-to-agent path on HTTP, where developers already are, and moves only the response path to a durable session. The agent essentially acknowledges the request over HTTP and then streams back over the session.<\/p>\n

Frameworks are baking in plug-in points for exactly this layer. Vercel<\/a> and TanStack both already expose transport abstractions for durable session providers. Ably\u2019s Vercel AI SDK plug-in adds multi-device session resumption, interruption support, and persistent storage.<\/p>\n

O\u2019Riordan doesn\u2019t want developers thinking in old-school pub\/sub terms either. \u201cI feel strongly that people should not have to think about pub\/sub.\u201d\u00a0<\/p>\n

\t<\/p>\n

\n\t\t\t\tYOUTUBE.COM\/THENEWSTACK\n\t\t\t<\/p>\n

\n\t\t\t\tTech moves fast, don’t miss an episode. Subscribe to our YouTube
\n\t\t\t\tchannel to stream all our podcasts, interviews, demos, and more.\n\t\t\t<\/p>\n

\t\t\t\tSUBSCRIBE<\/p>\n

\t<\/a><\/p>\n

Group
\n Created with Sketch.<\/p>\n

\t\t<\/p>\n

\t\t\t\t\t<\/p>\n

\n\t\t\t\t\t\t\tBefore joining The New Stack as its senior editor for AI, Frederic was the enterprise editor at TechCrunch, where he covered everything from the rise of the cloud and the earliest days of Kubernetes to the advent of quantum computing….\t\t\t\t\t\t<\/p>\n

\t\t\t\t\t\tRead more from Frederic Lardinois\t\t\t\t\t\t<\/p>\n

\t\t<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"For this episode of The New Stack Makers, we speak with Matthew O\u2019Riordan, CEO and co-founder of Ably.…\n","protected":false},"author":2,"featured_media":30683,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[19922,405,7537,3229,4898,2454],"class_list":{"0":"post-30682","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-agentic-ai","8":"tag-ably","9":"tag-ai-agents","10":"tag-artificial-intelligence-agents","11":"tag-podcast","12":"tag-post","13":"tag-video"},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/30682","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=30682"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/30682\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/30683"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=30682"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=30682"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=30682"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}