{"slug": "langgraph", "name": "LangGraph", "category": "ai", "category_name": "AI models and assistants", "description": "Open-source framework for building stateful, graph-based LLM agent workflows, from the LangChain team.", "recommenders": 7, "commenters": 17, "warners": 2, "score": 5.697, "rising": -2.29, "first": "2025-12-09", "latest": "2026-09-11", "why": "HN commenters use it to orchestrate agent systems as state machines, including in production, valuing its modularity, explicit control over logic, and features like time travel and human-in-the-loop interruptions.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 7, "neutral": 7, "compare": 1, "warn": 2}, "spark": [0, 2, 1, 0, 0, 1, 1, 2, 0, 0, 0, 0], "listed": true, "loved": ["Agent orchestration", "Control and reliability"], "caveat": "Reliability effort", "short_quote": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions)", "icon": true, "domain": null, "github_repo": null, "official_url": "https://www.langchain.com/langgraph", "pricing": "open source", "aliases": ["LangGraph", "langgraph"], "verdict": [{"text": "HN commenters use it to orchestrate agent systems as state machines, including in production,", "cites": [{"id": 739303808, "item_id": 46206488, "url": "https://news.ycombinator.com/item?id=46206488", "by": "arizen", "date": "2025-12-09", "stance": "recommend", "quote": "We use LangGraph to implement the State Machine pattern rather than a linear Chain.", "excerpt": "Hi HN, I wanted to share a reference implementation I architected for moving AI Agents from local prototypes to production services. The Context: It is relatively easy to get an agent working on a local machine where you can watch the terminal output and restart it if it gets stuck. However, the architecture often breaks down when moving to a headless, hosted environment where the agent needs to handle loops, persist", "context": "", "use_case": "agent state management", "switched_from": "", "story": "Show HN: Open-source reference architecture for AI Agents (LangGraph, Pydantic)"}, {"id": 773819985, "item_id": 48363749, "url": "https://news.ycombinator.com/item?id=48363749", "by": "coresmith", "date": "2026-06-01", "stance": "recommend", "quote": "CoreSmith is built on LangGraph.", "excerpt": "We built an agentic system that converts prompts into silicon using LangGraph. You provide a specification of the ASIC that you want, and a team of agents architects, designs, and tapes out a GDS. You can inspect the waveform to confirm pixels are moving, and you can look at the GDS to confirm that everything is routed: http://coresmith.ai/demo. CoreSmith iterates on the design until foundry DRC checks are clean \u2014 th", "context": "", "use_case": "orchestrating design agents", "switched_from": "", "story": "Prompt to Silicon with LangGraph"}, {"id": 763043600, "item_id": 47690225, "url": "https://news.ycombinator.com/item?id=47690225", "by": "Jignesh_1313", "date": "2026-04-08", "stance": "recommend", "quote": "LangGraph for the agent orchestration, async pipelines throughout.", "excerpt": "The multi-agent + RAG combo is what I'm living in right now. At AmberFlux EdgeAI I'm building SHEELA \u2014 an intelligence layer that crawls, scores, and runs controlled self-improvement loops on top of a FastAPI/PostgreSQL infrastructure backbone I architected from scratch. LangGraph for the agent orchestration, async pipelines throughout. But the project that feels most relevant to what you're describing is something I", "context": "final-year engineer building AI projects", "use_case": "agent orchestration", "switched_from": "", "story": "Ask HN: Who is hiring? (April 2026)"}, {"id": 747851792, "item_id": 46740737, "url": "https://news.ycombinator.com/item?id=46740737", "by": "dragonmage65535", "date": "2026-01-23", "stance": "neutral", "quote": "we use langgraph in production", "excerpt": "Yeah, I'm also exploring the same, we use langgraph in production, but seeing how much more powerful claude code is with little guidance, I just have more trust on it, but the library right now is a library with a subprocess (claude code) and the tools also need to be served via MCP, not exactly light weight, and also it would be good to have more visibility. Did you find any good resources on this?", "context": "", "use_case": "agents in production", "switched_from": "", "story": "Ask HN: Anyone using Claude Agent SDK in production?"}]}, {"text": " valuing its modularity, explicit control over logic, and features like time travel and human-in-the-loop interruptions.", "cites": [{"id": 776298385, "item_id": 48518649, "url": "https://news.ycombinator.com/item?id=48518649", "by": "mfn-throw", "date": "2026-06-13", "stance": "recommend", "quote": "For the agentic architecture we picked LangGraph \u2014 the modularity makes components easier to test.", "excerpt": "Hi HN! Excited to share something my friend and I have been building over many weekends. I'm a physicist turned software engineer; my buddy went from consulting to tech. We met at a college where the popular path after graduation was consulting or finance (over a decade ago) so we have a lot of friends in the space. We saw firsthand how much of a consultant/analyst's life is spent in PowerPoint: nudging boxes around,", "context": "", "use_case": "agent architecture", "switched_from": "", "story": "Show HN: Brightdeck \u2013 an OOXML-compatible AI presentation maker"}, {"id": 747910944, "item_id": 46744434, "url": "https://news.ycombinator.com/item?id=46744434", "by": "storystarling", "date": "2026-01-24", "stance": "neutral", "quote": "having that text state to pass into something like LangGraph is the only way to reliably control the logic", "excerpt": "I suspect the glued pipeline is going to remain dominant for a while, mostly because the intermediate text layer is structural, not just a byproduct. If you drop the text for a pure E2E model, you suddenly lose the ability to easily inject RAG context or handle complex tool use. I've been building some agent workflows recently and having that text state to pass into something like LangGraph is the only way to reliabl", "context": "building agent workflows", "use_case": "agent workflows", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 769207024, "item_id": 48075439, "url": "https://news.ycombinator.com/item?id=48075439", "by": "cucho", "date": "2026-05-09", "stance": "recommend", "quote": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions)", "excerpt": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions), flexibility on the paradigm.", "context": "", "use_case": "multi-agent platform", "switched_from": "", "story": "Ask HN: What is the underlying stack behind multi-agent platforms?"}, {"id": 739821008, "item_id": 46238813, "url": "https://news.ycombinator.com/item?id=46238813", "by": "solarkraft", "date": "2025-12-11", "stance": "recommend", "quote": "it seems to provide a lot of infrastructure for highly reliable and interactive agents", "excerpt": "This looks really cool for DIYing workflows, especially since you seem to have a very useful selection of tools! Did you build your own agent engine? Why not LangGraph? Say I was building a general agentic chat app with LangGraph in the backend (as it seems to provide a lot of infrastructure for highly reliable and interactive agents, all the way up to a protocol usable by UIs, plus a decent ecosystem, making it very", "context": "building an agentic chat app", "use_case": "agentic chat app backend", "switched_from": "", "story": "Show HN: Sim \u2013 Apache-2.0 n8n alternative"}]}, {"text": " The most common caveat is that agents built this way can be annoying to make reliable,", "cites": [{"id": 781973810, "item_id": 48873363, "url": "https://news.ycombinator.com/item?id=48873363", "by": "reveriedev", "date": "2026-07-11", "stance": "warn", "quote": "Getting those agent systems to behave reliably was often annoying and tricky.", "excerpt": "Back in the medieval age of 2025, I used frameworks like PydanticAI and LangGraph to build agents. Getting those agent systems to behave reliably was often annoying and tricky. Claude Code and Codex have since become capable general-purpose agents. They each have an SDK, but those SDKs expose different interfaces and inherit their respective platform\u2019s quirks. While building CodeAlmanac, I wanted one way to work with", "context": "", "use_case": "building agents", "switched_from": "", "story": null}]}, {"text": " and some now prefer Claude Code or plain generated scripts over a dedicated agent framework.", "cites": [{"id": 782760720, "item_id": 48922545, "url": "https://news.ycombinator.com/item?id=48922545", "by": "brainless", "date": "2026-07-15", "stance": "warn", "quote": "I usually stay away from systems which already make it seem as if building agents is a ritual.", "excerpt": "It is not blasphemy if langgraph is trying to do that. As I understand langgraph manages orchestration in custom built agents. I usually stay away from systems which already make it seem as if building agents is a ritual. What I am saying is the opposite - use Claude Code or whatever else - generate actual \"programs\". Basically scripts. We have tons of ways for \"programs\" to interact with each other. Then have clearl", "context": "", "use_case": "agent orchestration", "switched_from": "", "story": "Towards a Harness That Can Do Anything"}, {"id": 747851792, "item_id": 46740737, "url": "https://news.ycombinator.com/item?id=46740737", "by": "dragonmage65535", "date": "2026-01-23", "stance": "neutral", "quote": "we use langgraph in production", "excerpt": "Yeah, I'm also exploring the same, we use langgraph in production, but seeing how much more powerful claude code is with little guidance, I just have more trust on it, but the library right now is a library with a subprocess (claude code) and the tools also need to be served via MCP, not exactly light weight, and also it would be good to have more visibility. Did you find any good resources on this?", "context": "", "use_case": "agents in production", "switched_from": "", "story": "Ask HN: Anyone using Claude Agent SDK in production?"}]}], "praise": [{"label": "Agent orchestration", "summary": "Commenters build multi-agent and multi-stage systems on it as the orchestration layer.", "quotes": [{"id": 773819985, "item_id": 48363749, "url": "https://news.ycombinator.com/item?id=48363749", "by": "coresmith", "date": "2026-06-01", "stance": "recommend", "quote": "CoreSmith is built on LangGraph.", "excerpt": "We built an agentic system that converts prompts into silicon using LangGraph. You provide a specification of the ASIC that you want, and a team of agents architects, designs, and tapes out a GDS. You can inspect the waveform to confirm pixels are moving, and you can look at the GDS to confirm that everything is routed: http://coresmith.ai/demo. CoreSmith iterates on the design until foundry DRC checks are clean \u2014 th", "context": "", "use_case": "orchestrating design agents", "switched_from": "", "story": "Prompt to Silicon with LangGraph"}, {"id": 763043600, "item_id": 47690225, "url": "https://news.ycombinator.com/item?id=47690225", "by": "Jignesh_1313", "date": "2026-04-08", "stance": "recommend", "quote": "LangGraph for the agent orchestration, async pipelines throughout.", "excerpt": "The multi-agent + RAG combo is what I'm living in right now. At AmberFlux EdgeAI I'm building SHEELA \u2014 an intelligence layer that crawls, scores, and runs controlled self-improvement loops on top of a FastAPI/PostgreSQL infrastructure backbone I architected from scratch. LangGraph for the agent orchestration, async pipelines throughout. But the project that feels most relevant to what you're describing is something I", "context": "final-year engineer building AI projects", "use_case": "agent orchestration", "switched_from": "", "story": "Ask HN: Who is hiring? (April 2026)"}, {"id": 739303808, "item_id": 46206488, "url": "https://news.ycombinator.com/item?id=46206488", "by": "arizen", "date": "2025-12-09", "stance": "recommend", "quote": "We use LangGraph to implement the State Machine pattern rather than a linear Chain.", "excerpt": "Hi HN, I wanted to share a reference implementation I architected for moving AI Agents from local prototypes to production services. The Context: It is relatively easy to get an agent working on a local machine where you can watch the terminal output and restart it if it gets stuck. However, the architecture often breaks down when moving to a headless, hosted environment where the agent needs to handle loops, persist", "context": "", "use_case": "agent state management", "switched_from": "", "story": "Show HN: Open-source reference architecture for AI Agents (LangGraph, Pydantic)"}]}, {"label": "Control and reliability", "summary": "Passing explicit state through the graph is seen as a reliable way to control agent logic.", "quotes": [{"id": 747910944, "item_id": 46744434, "url": "https://news.ycombinator.com/item?id=46744434", "by": "storystarling", "date": "2026-01-24", "stance": "neutral", "quote": "having that text state to pass into something like LangGraph is the only way to reliably control the logic", "excerpt": "I suspect the glued pipeline is going to remain dominant for a while, mostly because the intermediate text layer is structural, not just a byproduct. If you drop the text for a pure E2E model, you suddenly lose the ability to easily inject RAG context or handle complex tool use. I've been building some agent workflows recently and having that text state to pass into something like LangGraph is the only way to reliabl", "context": "building agent workflows", "use_case": "agent workflows", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 739821008, "item_id": 46238813, "url": "https://news.ycombinator.com/item?id=46238813", "by": "solarkraft", "date": "2025-12-11", "stance": "recommend", "quote": "it seems to provide a lot of infrastructure for highly reliable and interactive agents", "excerpt": "This looks really cool for DIYing workflows, especially since you seem to have a very useful selection of tools! Did you build your own agent engine? Why not LangGraph? Say I was building a general agentic chat app with LangGraph in the backend (as it seems to provide a lot of infrastructure for highly reliable and interactive agents, all the way up to a protocol usable by UIs, plus a decent ecosystem, making it very", "context": "building an agentic chat app", "use_case": "agentic chat app backend", "switched_from": "", "story": "Show HN: Sim \u2013 Apache-2.0 n8n alternative"}, {"id": 747709169, "item_id": 46731823, "url": "https://news.ycombinator.com/item?id=46731823", "by": "storystarling", "date": "2026-01-23", "stance": "recommend", "quote": "I switched from a similar parallelized setup to LangGraph precisely because the merge conflicts and redundant reasoning steps were killing my margins.", "excerpt": "I'm not sure \"wasted work is cheap\" really holds up when you look at the token unit economics for Claude, especially if you're hitting the API with full context for every parallel agent. I switched from a similar parallelized setup to LangGraph precisely because the merge conflicts and redundant reasoning steps were killing my margins. How are you handling the git conflicts without a coordinator?", "context": "", "use_case": "agent orchestration", "switched_from": "parallelized multi-agent setup", "story": "Multiclaude \u2013 Lightweight Multiagent Orchestrator"}]}, {"label": "Low-level features", "summary": "Offers time travel, human-in-the-loop interruptions, and flexibility on the paradigm.", "quotes": [{"id": 769207024, "item_id": 48075439, "url": "https://news.ycombinator.com/item?id=48075439", "by": "cucho", "date": "2026-05-09", "stance": "recommend", "quote": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions)", "excerpt": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions), flexibility on the paradigm.", "context": "", "use_case": "multi-agent platform", "switched_from": "", "story": "Ask HN: What is the underlying stack behind multi-agent platforms?"}]}, {"label": "Modularity", "summary": "Its modular design makes individual components easier to test.", "quotes": [{"id": 776298385, "item_id": 48518649, "url": "https://news.ycombinator.com/item?id=48518649", "by": "mfn-throw", "date": "2026-06-13", "stance": "recommend", "quote": "For the agentic architecture we picked LangGraph \u2014 the modularity makes components easier to test.", "excerpt": "Hi HN! Excited to share something my friend and I have been building over many weekends. I'm a physicist turned software engineer; my buddy went from consulting to tech. We met at a college where the popular path after graduation was consulting or finance (over a decade ago) so we have a lot of friends in the space. We saw firsthand how much of a consultant/analyst's life is spent in PowerPoint: nudging boxes around,", "context": "", "use_case": "agent architecture", "switched_from": "", "story": "Show HN: Brightdeck \u2013 an OOXML-compatible AI presentation maker"}]}, {"label": "Model agnostic", "summary": "Described as model agnostic.", "quotes": [{"id": 750685572, "item_id": 46917848, "url": "https://news.ycombinator.com/item?id=46917848", "by": "moomooskycow", "date": "2026-02-06", "stance": "neutral", "quote": "I think the OpenAI Agents SDK is more model agnostic, LangGraph too.", "excerpt": "Seems like there are three things you want here: 1. Internal agent 2. Product features 3. Customer-facing agent This could all be served by one agent, but might not be as effective as having one focused agent for each need. It's not hard to build shared context for each agent to use, and it's (generally) easier to configure somewhat domain-specific agents than one superagent. Regarding hosting your own company agent:", "context": "", "use_case": "model-agnostic agents", "switched_from": "", "story": "Ask HN: Best option for hosted agent in 2026?"}]}], "caveats": [{"label": "Reliability effort", "summary": "One commenter found getting LangGraph agents to behave reliably annoying and tricky.", "quotes": [{"id": 781973810, "item_id": 48873363, "url": "https://news.ycombinator.com/item?id=48873363", "by": "reveriedev", "date": "2026-07-11", "stance": "warn", "quote": "Getting those agent systems to behave reliably was often annoying and tricky.", "excerpt": "Back in the medieval age of 2025, I used frameworks like PydanticAI and LangGraph to build agents. Getting those agent systems to behave reliably was often annoying and tricky. Claude Code and Codex have since become capable general-purpose agents. They each have an SDK, but those SDKs expose different interfaces and inherit their respective platform\u2019s quirks. While building CodeAlmanac, I wanted one way to work with", "context": "", "use_case": "building agents", "switched_from": "", "story": null}]}, {"label": "Coding agents alternative", "summary": "Some commenters trust Claude Code more or prefer generating plain scripts instead of using an agent framework.", "quotes": [{"id": 747851792, "item_id": 46740737, "url": "https://news.ycombinator.com/item?id=46740737", "by": "dragonmage65535", "date": "2026-01-23", "stance": "neutral", "quote": "we use langgraph in production", "excerpt": "Yeah, I'm also exploring the same, we use langgraph in production, but seeing how much more powerful claude code is with little guidance, I just have more trust on it, but the library right now is a library with a subprocess (claude code) and the tools also need to be served via MCP, not exactly light weight, and also it would be good to have more visibility. Did you find any good resources on this?", "context": "", "use_case": "agents in production", "switched_from": "", "story": "Ask HN: Anyone using Claude Agent SDK in production?"}, {"id": 782760720, "item_id": 48922545, "url": "https://news.ycombinator.com/item?id=48922545", "by": "brainless", "date": "2026-07-15", "stance": "warn", "quote": "I usually stay away from systems which already make it seem as if building agents is a ritual.", "excerpt": "It is not blasphemy if langgraph is trying to do that. As I understand langgraph manages orchestration in custom built agents. I usually stay away from systems which already make it seem as if building agents is a ritual. What I am saying is the opposite - use Claude Code or whatever else - generate actual \"programs\". Basically scripts. We have tons of ways for \"programs\" to interact with each other. Then have clearl", "context": "", "use_case": "agent orchestration", "switched_from": "", "story": "Towards a Harness That Can Do Anything"}]}, {"label": "Low level", "summary": "Described as quite low level.", "quotes": [{"id": 769207024, "item_id": 48075439, "url": "https://news.ycombinator.com/item?id=48075439", "by": "cucho", "date": "2026-05-09", "stance": "recommend", "quote": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions)", "excerpt": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions), flexibility on the paradigm.", "context": "", "use_case": "multi-agent platform", "switched_from": "", "story": "Ask HN: What is the underlying stack behind multi-agent platforms?"}]}], "who": [{"by": "jabrotapu", "context": "freelancer building on-prem GPU clusters", "url": "https://news.ycombinator.com/item?id=49661006", "date": "2026-09-11"}, {"by": "emanuelecz", "context": "self-taught AI engineer", "url": "https://news.ycombinator.com/item?id=49224408", "date": "2026-08-08"}, {"by": "ilaksh", "context": "freelance developer on Upwork", "url": "https://news.ycombinator.com/item?id=49160257", "date": "2026-08-03"}, {"by": "Jignesh_1313", "context": "final-year engineer building AI projects", "url": "https://news.ycombinator.com/item?id=47690225", "date": "2026-04-08"}, {"by": "seudxs", "context": "solo project", "url": "https://news.ycombinator.com/item?id=47393215", "date": "2026-03-15"}, {"by": "storystarling", "context": "building agent workflows", "url": "https://news.ycombinator.com/item?id=46744434", "date": "2026-01-24"}, {"by": "solarkraft", "context": "building an agentic chat app", "url": "https://news.ycombinator.com/item?id=46238813", "date": "2025-12-11"}], "tags": ["for multi-agent systems", "for production agents", "for human-in-the-loop agents", "model agnostic"], "use_cases": ["agent orchestration", "building agents", "agent workflows", "building the app", "agents", "client-requested agent framework"], "compared": [{"slug": "claude-code", "name": "Claude Code", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "postgresql", "name": "PostgreSQL", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "claude", "name": "Claude", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "fastapi", "name": "FastAPI", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "claude-agent-sdk", "name": "Claude Agent SDK", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "cursor", "name": "Cursor", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "docker", "name": "Docker", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=49522905", "title": "Ask HN: Freelancer? Seeking freelancer? (September 2026)", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49426349", "title": "Elon House", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49156682", "title": "Ask HN: Who wants to be hired? (August 2026)", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49150470", "title": "Qwen3.8-Max: A New Bar for Coding and Cowork", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48921077", "title": "Towards a Harness That Can Do Anything", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48873362", "title": "item 48873362", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {"recommend": 2}, "2026-01": {"recommend": 1, "neutral": 1}, "2026-02": {"neutral": 1}, "2026-03": {"neutral": 1}, "2026-04": {"recommend": 1, "compare": 1}, "2026-05": {"recommend": 1}, "2026-06": {"recommend": 2}, "2026-07": {"warn": 2}, "2026-08": {"neutral": 3}, "2026-09": {"neutral": 1}, "2026-10": {}}, "quotes": [{"id": 794576109, "item_id": 49661006, "url": "https://news.ycombinator.com/item?id=49661006", "by": "jabrotapu", "date": "2026-09-11", "stance": "neutral", "quote": "Agents: LangGraph, LangChain", "excerpt": "SEEKING WORK | India, remote | Part-time | Overlap with US, UK, EU hours I build on-prem GPU clusters and private AI apps that run entirely on the client's own hardware, so their data never leaves the building. About 10 years full-stack, the last few on GPU infrastructure. Currently running a live production cluster: 700-800 users sharing a small physical GPU pool. GPU-backed JupyterHub, single sign-on, per-user quot", "context": "freelancer building on-prem GPU clusters", "use_case": "building agents", "switched_from": "", "story": "Ask HN: Freelancer? Seeking freelancer? (September 2026)"}, {"id": 790821601, "item_id": 49426350, "url": "https://news.ycombinator.com/item?id=49426350", "by": "nicolasmery", "date": "2026-08-24", "stance": "neutral", "quote": "I used langgraph + ws + all in cloudflare + 2 llm providers", "excerpt": "Have you ever asked an AI to tell you a joke? They\u2019re so bad haha. This is a small experiment along those lines https://www.elon-house.com I built it a year ago, but decided to finally publish it after getting some great feedback. I used langgraph + ws + all in cloudflare + 2 llm providers", "context": "", "use_case": "building the app", "switched_from": "", "story": "Elon House"}, {"id": 787590530, "item_id": 49224408, "url": "https://news.ycombinator.com/item?id=49224408", "by": "emanuelecz", "date": "2026-08-08", "stance": "neutral", "quote": "Python, FastAPI, LangGraph, MCP, RAG, Postgres/pgvector, Redis, Celery, Docker", "excerpt": "Location: Gipuzkoa, Spain (CET) Remote: Yes, remote only Willing to relocate: No Technologies: Python, FastAPI, LangGraph, MCP, RAG, Postgres/pgvector, Redis, Celery, Docker, LLM evals, TypeScript, Next.js R\u00e9sum\u00e9/CV: https://www.emanuelecz.dev/emanuele-czofei-cv.pdf Email: emanuelecz06@gmail.com Self-taught AI engineer, two years building, four months at a startup. Strongest at the unglamorous half: retrieval that re", "context": "self-taught AI engineer", "use_case": "agents", "switched_from": "", "story": "Ask HN: Who wants to be hired? (August 2026)"}, {"id": 786564114, "item_id": 49160257, "url": "https://news.ycombinator.com/item?id=49160257", "by": "ilaksh", "date": "2026-08-03", "stance": "neutral", "quote": "someone who wanted me to start over in LangGraph", "excerpt": "Not sure what you are trying to say but thanks for reminding me about qm, it has some cool features. My MindRoot framework had some of qm's main features awhile ago though. I do try to use it to build solutions by customizing rather than from scratch when possible. But a lot of clients or potential clients don't like that idea even though it's MIT. They want some that has clearly been invented by them as a new progra", "context": "freelance developer on Upwork", "use_case": "client-requested agent framework", "switched_from": "", "story": "Qwen3.8-Max: A New Bar for Coding and Cowork"}, {"id": 782760720, "item_id": 48922545, "url": "https://news.ycombinator.com/item?id=48922545", "by": "brainless", "date": "2026-07-15", "stance": "warn", "quote": "I usually stay away from systems which already make it seem as if building agents is a ritual.", "excerpt": "It is not blasphemy if langgraph is trying to do that. As I understand langgraph manages orchestration in custom built agents. I usually stay away from systems which already make it seem as if building agents is a ritual. What I am saying is the opposite - use Claude Code or whatever else - generate actual \"programs\". Basically scripts. We have tons of ways for \"programs\" to interact with each other. Then have clearl", "context": "", "use_case": "agent orchestration", "switched_from": "", "story": "Towards a Harness That Can Do Anything"}, {"id": 781973810, "item_id": 48873363, "url": "https://news.ycombinator.com/item?id=48873363", "by": "reveriedev", "date": "2026-07-11", "stance": "warn", "quote": "Getting those agent systems to behave reliably was often annoying and tricky.", "excerpt": "Back in the medieval age of 2025, I used frameworks like PydanticAI and LangGraph to build agents. Getting those agent systems to behave reliably was often annoying and tricky. Claude Code and Codex have since become capable general-purpose agents. They each have an SDK, but those SDKs expose different interfaces and inherit their respective platform\u2019s quirks. While building CodeAlmanac, I wanted one way to work with", "context": "", "use_case": "building agents", "switched_from": "", "story": null}, {"id": 776298385, "item_id": 48518649, "url": "https://news.ycombinator.com/item?id=48518649", "by": "mfn-throw", "date": "2026-06-13", "stance": "recommend", "quote": "For the agentic architecture we picked LangGraph \u2014 the modularity makes components easier to test.", "excerpt": "Hi HN! Excited to share something my friend and I have been building over many weekends. I'm a physicist turned software engineer; my buddy went from consulting to tech. We met at a college where the popular path after graduation was consulting or finance (over a decade ago) so we have a lot of friends in the space. We saw firsthand how much of a consultant/analyst's life is spent in PowerPoint: nudging boxes around,", "context": "", "use_case": "agent architecture", "switched_from": "", "story": "Show HN: Brightdeck \u2013 an OOXML-compatible AI presentation maker"}, {"id": 773819985, "item_id": 48363749, "url": "https://news.ycombinator.com/item?id=48363749", "by": "coresmith", "date": "2026-06-01", "stance": "recommend", "quote": "CoreSmith is built on LangGraph.", "excerpt": "We built an agentic system that converts prompts into silicon using LangGraph. You provide a specification of the ASIC that you want, and a team of agents architects, designs, and tapes out a GDS. You can inspect the waveform to confirm pixels are moving, and you can look at the GDS to confirm that everything is routed: http://coresmith.ai/demo. CoreSmith iterates on the design until foundry DRC checks are clean \u2014 th", "context": "", "use_case": "orchestrating design agents", "switched_from": "", "story": "Prompt to Silicon with LangGraph"}, {"id": 769207024, "item_id": 48075439, "url": "https://news.ycombinator.com/item?id=48075439", "by": "cucho", "date": "2026-05-09", "stance": "recommend", "quote": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions)", "excerpt": "We are using LangGraph, which is quite low level, but has useful features like time travel, human in the loop capabilities (interruptions), flexibility on the paradigm.", "context": "", "use_case": "multi-agent platform", "switched_from": "", "story": "Ask HN: What is the underlying stack behind multi-agent platforms?"}, {"id": 764609666, "item_id": 47788104, "url": "https://news.ycombinator.com/item?id=47788104", "by": "myui", "date": "2026-04-15", "stance": "compare", "quote": "as an alternative to LangGraph", "excerpt": "Looks interesting and quite polished overall. I\u2019m also working on an agentic workflow engine (graflow.ai) as an alternative to LangGraph, so this caught my attention. The sandboxing layer in particular looks well thought out and solid. One question I had: in the demo, the agent builds a browser game. But that kind of task seems achievable even with vanilla Opus (plus Claude Skills / tool use), so I struggled a bit to", "context": "", "use_case": "agent workflows", "switched_from": "", "story": "Show HN: Agentfab \u2013 A Distributed Agentic Platform"}, {"id": 763043600, "item_id": 47690225, "url": "https://news.ycombinator.com/item?id=47690225", "by": "Jignesh_1313", "date": "2026-04-08", "stance": "recommend", "quote": "LangGraph for the agent orchestration, async pipelines throughout.", "excerpt": "The multi-agent + RAG combo is what I'm living in right now. At AmberFlux EdgeAI I'm building SHEELA \u2014 an intelligence layer that crawls, scores, and runs controlled self-improvement loops on top of a FastAPI/PostgreSQL infrastructure backbone I architected from scratch. LangGraph for the agent orchestration, async pipelines throughout. But the project that feels most relevant to what you're describing is something I", "context": "final-year engineer building AI projects", "use_case": "agent orchestration", "switched_from": "", "story": "Ask HN: Who is hiring? (April 2026)"}, {"id": 758291442, "item_id": 47393215, "url": "https://news.ycombinator.com/item?id=47393215", "by": "seudxs", "date": "2026-03-15", "stance": "neutral", "quote": "I used AI agent for my last langgraph solo project", "excerpt": "I started AI-assisted coding quite a while ago with \"query for code to copy and paste\" approach which was slow but it dramatically shifts when the LLMs are used as agents that are just AI that have access to certain things like your project's source codes, internet and some other technical docs that refines them. You can simply instruct it to change snippet of codes by mentioning them with their actions inside the ch", "context": "solo project", "use_case": "agent project", "switched_from": "", "story": "Ask HN: How is AI-assisted coding going for you professionally?"}, {"id": 750685572, "item_id": 46917848, "url": "https://news.ycombinator.com/item?id=46917848", "by": "moomooskycow", "date": "2026-02-06", "stance": "neutral", "quote": "I think the OpenAI Agents SDK is more model agnostic, LangGraph too.", "excerpt": "Seems like there are three things you want here: 1. Internal agent 2. Product features 3. Customer-facing agent This could all be served by one agent, but might not be as effective as having one focused agent for each need. It's not hard to build shared context for each agent to use, and it's (generally) easier to configure somewhat domain-specific agents than one superagent. Regarding hosting your own company agent:", "context": "", "use_case": "model-agnostic agents", "switched_from": "", "story": "Ask HN: Best option for hosted agent in 2026?"}, {"id": 747910944, "item_id": 46744434, "url": "https://news.ycombinator.com/item?id=46744434", "by": "storystarling", "date": "2026-01-24", "stance": "neutral", "quote": "having that text state to pass into something like LangGraph is the only way to reliably control the logic", "excerpt": "I suspect the glued pipeline is going to remain dominant for a while, mostly because the intermediate text layer is structural, not just a byproduct. If you drop the text for a pure E2E model, you suddenly lose the ability to easily inject RAG context or handle complex tool use. I've been building some agent workflows recently and having that text state to pass into something like LangGraph is the only way to reliabl", "context": "building agent workflows", "use_case": "agent workflows", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 747851792, "item_id": 46740737, "url": "https://news.ycombinator.com/item?id=46740737", "by": "dragonmage65535", "date": "2026-01-23", "stance": "neutral", "quote": "we use langgraph in production", "excerpt": "Yeah, I'm also exploring the same, we use langgraph in production, but seeing how much more powerful claude code is with little guidance, I just have more trust on it, but the library right now is a library with a subprocess (claude code) and the tools also need to be served via MCP, not exactly light weight, and also it would be good to have more visibility. Did you find any good resources on this?", "context": "", "use_case": "agents in production", "switched_from": "", "story": "Ask HN: Anyone using Claude Agent SDK in production?"}, {"id": 747709169, "item_id": 46731823, "url": "https://news.ycombinator.com/item?id=46731823", "by": "storystarling", "date": "2026-01-23", "stance": "recommend", "quote": "I switched from a similar parallelized setup to LangGraph precisely because the merge conflicts and redundant reasoning steps were killing my margins.", "excerpt": "I'm not sure \"wasted work is cheap\" really holds up when you look at the token unit economics for Claude, especially if you're hitting the API with full context for every parallel agent. I switched from a similar parallelized setup to LangGraph precisely because the merge conflicts and redundant reasoning steps were killing my margins. How are you handling the git conflicts without a coordinator?", "context": "", "use_case": "agent orchestration", "switched_from": "parallelized multi-agent setup", "story": "Multiclaude \u2013 Lightweight Multiagent Orchestrator"}, {"id": 739821008, "item_id": 46238813, "url": "https://news.ycombinator.com/item?id=46238813", "by": "solarkraft", "date": "2025-12-11", "stance": "recommend", "quote": "it seems to provide a lot of infrastructure for highly reliable and interactive agents", "excerpt": "This looks really cool for DIYing workflows, especially since you seem to have a very useful selection of tools! Did you build your own agent engine? Why not LangGraph? Say I was building a general agentic chat app with LangGraph in the backend (as it seems to provide a lot of infrastructure for highly reliable and interactive agents, all the way up to a protocol usable by UIs, plus a decent ecosystem, making it very", "context": "building an agentic chat app", "use_case": "agentic chat app backend", "switched_from": "", "story": "Show HN: Sim \u2013 Apache-2.0 n8n alternative"}, {"id": 739303808, "item_id": 46206488, "url": "https://news.ycombinator.com/item?id=46206488", "by": "arizen", "date": "2025-12-09", "stance": "recommend", "quote": "We use LangGraph to implement the State Machine pattern rather than a linear Chain.", "excerpt": "Hi HN, I wanted to share a reference implementation I architected for moving AI Agents from local prototypes to production services. The Context: It is relatively easy to get an agent working on a local machine where you can watch the terminal output and restart it if it gets stuck. However, the architecture often breaks down when moving to a headless, hosted environment where the agent needs to handle loops, persist", "context": "", "use_case": "agent state management", "switched_from": "", "story": "Show HN: Open-source reference architecture for AI Agents (LangGraph, Pydantic)"}]}