[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-open-source-agent-stacks-seven-layers-2026-en":3,"article-related-open-source-agent-stacks-seven-layers-2026-en":31,"series-industry-4abfb46f-b476-4837-abe0-823deb9bef92":78},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":23,"views":27,"created_at":28,"published_at":29,"topic_cluster_id":30},"4abfb46f-b476-4837-abe0-823deb9bef92","open-source-agent-stacks-seven-layers-2026-en","Open source agent stacks split into seven layers in 2026","\u003Cp data-speakable=\"summary\">O’Reilly’s 2026 guide breaks \u003Ca href=\"\u002Ftag\u002Fagent\">agent\u003C\u002Fa> tooling into seven production layers.\u003C\u002Fp>\u003Cp>\u003Ca href=\"\u002Fnews\u002Fopen-source-ai-agent-frameworks-compared-langfuse-en\">Open source agent\u003C\u002Fa> stacks in 2026 are no longer one big framework problem. Paolo Perrone’s O’Reilly Radar piece, published July 14, 2026, argues that teams should choose tools by layer, because the thing that fails first in production is rarely the same thing that impressed everyone in the demo.\u003C\u002Fp>\u003Cp>The article’s core warning is simple: \u003Ca href=\"\u002Ftag\u002Fbenchmark\">benchmark\u003C\u002Fa> wins can hide ugly tradeoffs. One memory framework may crush \u003Ca href=\"https:\u002F\u002Flilys.ai\u002Fnotes\u002F1542\" target=\"_blank\" rel=\"noopener\">LoCoMo\u003C\u002Fa>, the long-conversation memory benchmark, while running 340x heavier per conversation than the runner-up. That kind of gap changes architecture decisions fast.\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>Signal\u003C\u002Fth>\u003Cth>Value\u003C\u002Fth>\u003Cth>Why it matters\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>Publication date\u003C\u002Ftd>\u003Ctd>July 14, 2026\u003C\u002Ftd>\u003Ctd>Places the analysis in the current agent-tooling cycle\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Memory benchmark gap\u003C\u002Ftd>\u003Ctd>340x heavier\u003C\u002Ftd>\u003Ctd>Shows why scores alone can mislead teams\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>GitHub stars\u003C\u002Ftd>\u003Ctd>50,000+\u003C\u002Ftd>\u003Ctd>Signals fast adoption for the leading Python agent project\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Ch2>The seven-layer split is the real story\u003C\u002Fh2>\u003Cp>Perrone’s main point is that the agent stack has matured into seven layers: orchestration, memory, tool interface, browser or computer-use tools, coding agents, evals and observability, and \u003Ca href=\"\u002Ftag\u002Finference\">inference\u003C\u002Fa>. That split matters because each layer has a different bottleneck, and each bottleneck punishes a different kind of mistake.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784403178969-fbjv.png\" alt=\"Open source agent stacks split into seven layers in 2026\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>He frames the choice around four constraints: latency budget, audit trail, model portability, and language stack. That is a much better filter than asking which project has the flashiest demo video or the most active Discord.\u003C\u002Fp>\u003Cp>The article also draws a hard line between open source and open core. If production features like multitenant auth, replication, SSO, or audit logs live only in the hosted product, the repo is not the whole answer for a production team.\u003C\u002Fp>\u003Cul>\u003Cli>Latency budget decides how much time and token spend each turn can consume.\u003C\u002Fli>\u003Cli>Audit trail matters when every action needs to be traceable for compliance.\u003C\u002Fli>\u003Cli>Model portability matters when you do not want the agent locked to one provider.\u003C\u002Fli>\u003Cli>Language stack matters because Python and TypeScript teams do not want the same tooling.\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>Orchestration is where production pain starts\u003C\u002Fh2>\u003Cp>At the orchestration layer, the article puts \u003Ca href=\"https:\u002F\u002Flangchain-ai.github.io\u002Flanggraph\u002F\" target=\"_blank\" rel=\"noopener\">LangGraph\u003C\u002Fa> in the default slot for Python production work. The reason is not hype. It is state management: durable execution, checkpointing with \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flangchain-ai\u002Flanggraph\u002Ftree\u002Fmain\u002Flibs\u002Fcheckpoint-postgres\" target=\"_blank\" rel=\"noopener\">PostgresSaver\u003C\u002Fa>, time-travel debugging, and a model that maps cleanly onto audit-heavy systems.\u003C\u002Fp>\u003Cp>The author points out that companies such as \u003Ca href=\"https:\u002F\u002Fwww.klarna.com\u002F\" target=\"_blank\" rel=\"noopener\">Klarna\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Fwww.uber.com\u002F\" target=\"_blank\" rel=\"noopener\">Uber\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Fwww.linkedin.com\u002F\" target=\"_blank\" rel=\"noopener\">LinkedIn\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Fwww.jpmorgan.com\u002F\" target=\"_blank\" rel=\"noopener\">JPMorgan\u003C\u002Fa>, and \u003Ca href=\"https:\u002F\u002Freplit.com\u002F\" target=\"_blank\" rel=\"noopener\">Replit\u003C\u002Fa> appear in the framework’s verified enterprise list. That matters because production adoption is often the strongest signal that a framework has survived contact with real systems.\u003C\u002Fp>\u003Cp>Perrone is also blunt about the tradeoff: LangGraph is verbose. A simple two-agent flow still needs a state schema, nodes, edges, and compilation. If your workflow is just “call three tools sequentially,” that overhead can feel like using a freight train to move a backpack.\u003C\u002Fp>\u003Cblockquote>\u003Cp>“The best way to zero in on the constraint your system will hit first under load: latency budget, audit trail, model portability, or language stack.”\u003C\u002Fp>\u003Cfooter>Paolo Perrone, O’Reilly Radar, July 14, 2026\u003C\u002Ffooter>\u003C\u002Fblockquote>\u003Cp>That quote is the article in one sentence. It \u003Ca href=\"\u002Fnews\u002Fgrok-4-5-one-prompt-agent-work-en\">turns agent\u003C\u002Fa> selection from a feature checklist into an engineering constraint exercise.\u003C\u002Fp>\u003Ch2>The lighter tools win when the problem is smaller\u003C\u002Fh2>\u003Cp>The article does not crown one framework for everything. It gives each tool a job, and the jobs are different enough that the comparisons matter. \u003Ca href=\"https:\u002F\u002Fwww.crewai.com\u002F\" target=\"_blank\" rel=\"noopener\">CrewAI\u003C\u002Fa> gets the nod for lowest setup overhead. You define roles such as researcher, writer, and reviewer, then let the crew run with minimal ceremony.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784403173879-3rmg.png\" alt=\"Open source agent stacks split into seven layers in 2026\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>That simplicity comes with a cost. CrewAI does not resume crashed runs from the point of failure, error handling sits at the crew level rather than the node level, and the system does not keep an inspectable state schema that records what happened and when. For prototypes, that may be fine. For systems that need postmortems, it becomes a problem.\u003C\u002Fp>\u003Cp>\u003Ca href=\"https:\u002F\u002Fai.pydantic.dev\u002F\" target=\"_blank\" rel=\"noopener\">Pydantic AI\u003C\u002Fa> takes a different route. It treats every output as a typed Pydantic model, which makes validation and serialization much cleaner for single-loop agents that feed downstream services. The tradeoff is weaker multi-agent support than LangGraph or CrewAI.\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ca href=\"https:\u002F\u002Flangchain-ai.github.io\u002Flanggraph\u002F\" target=\"_blank\" rel=\"noopener\">LangGraph\u003C\u002Fa>: best for durable Python production flows\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fwww.crewai.com\u002F\" target=\"_blank\" rel=\"noopener\">CrewAI\u003C\u002Fa>: best for quick role-based prototypes\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fai.pydantic.dev\u002F\" target=\"_blank\" rel=\"noopener\">Pydantic AI\u003C\u002Fa>: best for typed single-agent outputs\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fmastra.ai\u002F\" target=\"_blank\" rel=\"noopener\">Mastra\u003C\u002Fa>: best for TypeScript teams building inside Next.js\u003C\u002Fli>\u003C\u002Ful>\u003Cp>For \u003Ca href=\"\u002Ftag\u002Ftypescript\">TypeScript\u003C\u002Fa> shops, \u003Ca href=\"https:\u002F\u002Fmastra.ai\u002F\" target=\"_blank\" rel=\"noopener\">Mastra\u003C\u002Fa> gets a strong mention because it bundles agents, workflows, \u003Ca href=\"\u002Ftag\u002Frag\">RAG\u003C\u002Fa>, and evals into one package. The article notes that the project comes from the ex-Gatsby founders and fits naturally into existing \u003Ca href=\"https:\u002F\u002Fnextjs.org\u002F\" target=\"_blank\" rel=\"noopener\">Next.js\u003C\u002Fa> apps without adding a Python sidecar.\u003C\u002Fp>\u003Cp>The comparison is practical rather than ideological. If your team is already all-in on TypeScript, forcing a Python stack just to satisfy a tooling trend makes little sense. If your team needs stateful retries, traceability, and a larger enterprise footprint, the Python path still looks better.\u003C\u002Fp>\u003Ch2>Benchmark numbers matter less than failure modes\u003C\u002Fh2>\u003Cp>The most useful part of the piece is how it treats benchmark numbers. A win on a leaderboard can still hide a cost that kills you later. The article’s 340x memory gap is the clearest example, but the same logic applies across the stack: browser tools fail on canvas-heavy sites, eval suites rot when they live in a Notion page, and inference choices can lock you into one vendor.\u003C\u002Fp>\u003Cp>That is why the piece keeps returning to production failure modes. The question is not “which project is best?” It is “which project fails in the least damaging way for my system?”\u003C\u002Fp>\u003Cp>That framing also explains why the article treats the open source toolkit as a collection of tradeoffs rather than a winner-take-all market. Some layers reward durability. Some reward speed of setup. Some reward typed outputs. Some reward ecosystem depth. Teams that ignore those differences end up rewriting schemas in week three.\u003C\u002Fp>\u003Ch2>What teams should do next\u003C\u002Fh2>\u003Cp>If you are building an agent in 2026, the article’s advice is to pick the first layer that will break under load and start there. For regulated workflows, that is often orchestration. For content or support bots, it may be memory or browser tooling. For product teams shipping inside a web app, it may be the language stack.\u003C\u002Fp>\u003Cp>The bigger takeaway is that the open source agent world has become more mature and more fragmented at the same time. That is good news for teams that know their constraints, and bad news for anyone hoping one framework will solve every problem.\u003C\u002Fp>\u003Cp>My read is that the next wave of adoption will reward boring reliability over demo polish. The teams that win will be the ones that can answer a simple question before they write code: which layer am I optimizing, and what am I willing to give up to make it work in production?\u003C\u002Fp>\u003C\u002Fcontent>","O’Reilly’s 2026 guide breaks agent tooling into seven layers and names the open source picks that survive production.","www.oreilly.com","https:\u002F\u002Fwww.oreilly.com\u002Fradar\u002Fthe-open-source-agent-toolkit-in-2026\u002F",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784403178969-fbjv.png","industry","en","2299155a-c8ca-40e1-9916-dd43f8f7a35f",[17,18,19,20,21,22],"open source agents","LangGraph","CrewAI","Pydantic AI","Mastra","agent tooling",[24,25,26],"The 2026 agent stack splits into seven layers, and each layer has different production constraints.","LangGraph leads Python production work because of durable execution and stateful debugging.","Benchmarks can hide huge runtime costs, so teams should choose tools by failure mode, not hype.",1,"2026-07-18T19:32:33.06165+00:00","2026-07-18T19:32:33.05+00:00","d1b328d5-54e2-4bd4-83cd-34d2c5c85089",{"tags":32,"relatedLang":37,"relatedPosts":41},[33,35],{"name":18,"slug":34},"langgraph",{"name":19,"slug":36},"crewai",{"id":15,"slug":38,"title":39,"language":40},"open-source-agent-stacks-seven-layers-2026-zh","2026 開源 agent 堆疊拆成七層","zh",[42,48,54,60,66,72],{"id":43,"slug":44,"title":45,"cover_image":46,"image_url":46,"created_at":47,"category":13},"0eb521a5-8e5c-4825-884f-18f5f045bca7","llm-benchmark-2026-38-real-tasks-15-models-en","15 LLMs on 38 tasks show routing beats one-model bets","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784422964229-9vo4.png","2026-07-19T01:02:19.445604+00:00",{"id":49,"slug":50,"title":51,"cover_image":52,"image_url":52,"created_at":53,"category":13},"530266f1-c175-4a07-aa6c-963662096df6","openai-staff-fund-rival-super-pac-en","$215,000 from OpenAI staff backs anti-AI PAC","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784419367270-a4em.png","2026-07-19T00:02:21.711542+00:00",{"id":55,"slug":56,"title":57,"cover_image":58,"image_url":58,"created_at":59,"category":13},"0785f6be-9cb9-436e-ad94-6035d7eeed2d","kimi-k3-model-hype-into-harness-work-en","Kimi K3 turns model hype into harness work","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784397790722-ckgl.png","2026-07-18T18:02:44.152791+00:00",{"id":61,"slug":62,"title":63,"cover_image":64,"image_url":64,"created_at":65,"category":13},"4b271e35-cbf2-4b31-a6af-5ea10e074e13","friends-season-7-episode-12-sitcom-machine-en","Friends Season 7, Episode 12 Is Still a Perfect Sitcom Machine","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784376190593-18ei.png","2026-07-18T12:02:39.200902+00:00",{"id":67,"slug":68,"title":69,"cover_image":70,"image_url":70,"created_at":71,"category":13},"109fb2c2-4d48-461a-a9b8-cf52ea8a7ac3","openai-bio-bounty-pay-more-for-jailbreaks-en","OpenAI should pay more for bio jailbreaks, not less","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784356374692-q6hu.png","2026-07-18T06:32:31.681069+00:00",{"id":73,"slug":74,"title":75,"cover_image":76,"image_url":76,"created_at":77,"category":13},"078103bc-b01d-4a64-948a-b94020f7c4b0","cloudflare-q2-2026-earnings-call-aug-6-en","Cloudflare sets its Q2 2026 earnings call for Aug. 6","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784316758954-u2zd.png","2026-07-17T19:32:20.369797+00:00",[79,84,89,94,99,104,109,114,119,124],{"id":80,"slug":81,"title":82,"created_at":83},"d35a1bd9-e709-412e-a2df-392df1dc572a","ai-impact-2026-developments-market-en","AI's Impact in 2026: Key Developments and Market Shifts","2026-03-25T16:20:33.205823+00:00",{"id":85,"slug":86,"title":87,"created_at":88},"5ed27921-5fd6-492e-8c59-78393bf37710","trumps-ai-legislative-framework-en","Trump's AI Legislative Framework: What's Inside?","2026-03-25T16:22:20.005325+00:00",{"id":90,"slug":91,"title":92,"created_at":93},"e454a642-f03c-4794-b185-5f651aebbaca","nvidia-gtc-2026-key-highlights-innovations-en","NVIDIA GTC 2026: Key Highlights and Innovations","2026-03-25T16:22:47.882615+00:00",{"id":95,"slug":96,"title":97,"created_at":98},"0ebb5b16-774a-4922-945d-5f2ce1df5a6d","claude-usage-diversifies-learning-curves-en","Claude Usage Diversifies, Learning Curves Emerge","2026-03-25T16:25:50.770376+00:00",{"id":100,"slug":101,"title":102,"created_at":103},"69934e86-2fc5-4280-8223-7b917a48ace8","openclaw-ai-commoditization-concerns-en","OpenClaw's Rise Raises Concerns of AI Model Commoditization","2026-03-25T16:26:30.582047+00:00",{"id":105,"slug":106,"title":107,"created_at":108},"b4b2575b-2ac8-46b2-b90e-ab1d7c060797","google-gemini-ai-rollout-2026-en","Google's Gemini AI Rollout Extended to 2026","2026-03-25T16:28:14.808842+00:00",{"id":110,"slug":111,"title":112,"created_at":113},"6e18bc65-42ae-4ad0-b564-67d7f66b979e","meta-llama4-fabricated-results-scandal-en","Meta's Llama 4 Scandal: Fabricated AI Test Results Unveiled","2026-03-25T16:29:15.482836+00:00",{"id":115,"slug":116,"title":117,"created_at":118},"bf888e9d-08be-4f47-996c-7b24b5ab3500","accenture-mistral-ai-deployment-en","Accenture and Mistral AI Team Up for AI Deployment","2026-03-25T16:31:01.894655+00:00",{"id":120,"slug":121,"title":122,"created_at":123},"5382b536-fad2-49c6-ac85-9eb2bae49f35","mistral-ai-high-stakes-2026-en","Mistral AI: Facing High Stakes in 2026","2026-03-25T16:31:39.941974+00:00",{"id":125,"slug":126,"title":127,"created_at":128},"9da3d2d6-b669-4971-ba1d-17fdb3548ed5","cursors-meteoric-rise-pressures-en","Cursor's Meteoric Rise Faces Industry Pressures","2026-03-25T16:32:21.899217+00:00"]