Knowledge catalogue
“Every enterprise needs a claw strategy.” How did @LangChain go from a weekend project to 1B+ downloads in 3 years? We sat down with CEO and co-founder Harrison Chase (@hwchase17) to talk deep agents,
It feels like agent harness evolution runs on two axes that usually get conflated. There’s the temporal axis: simplify as models improve, stripping components that compensated for limitations the new
.@ListenLabs Co-Founder & CTO @florian_jue on the 2 types of engineers who build great agents: 1/ Those who know what LLMs can and can't do, and can feel when something's off 2/ Product engineers who
really cool work by the Harvey team, excited to partner with them to push forward research on designing + understanding agents across Long Horizon Legal work the first peak I got at LAB was 'woah this
small workflow note that adds up. /staged-pr is a skill (via slash command) i run when wrapping up a PR. it takes my staged code changes and drafts a concise PR title & description, based on my prefer
Strong Opinions, Loosely Held on Agent + Harness Engineering: 1. You can outperform any default harness+model (including codex & claude code) on pretty much any Task by engineering the harness around
the deeper point: model choice and harness choice aren't independent variables. benchmarking a model without specifying the harness is like benchmarking a chip without specifying the compiler. the num
this deepagents deploy https://docs.langchain.com/oss/python/deepagents/deploy (or at least directionally where we want to take it) what's missing? give us feedback! can someone PLEASE launch OS claud
Harrison Chase expresses enthusiasm about collaborating with Harvey on their Long Horizon Legal Agent Benchmark, highlighting the value of developing industry-specific benchmarks for AI evaluation. Th
Harrison Chase expresses support for expanding AI benchmarking efforts and highlights appreciation for work being done by Harvey (likely Harvey AI, a legal AI company). The post indicates enthusiasm f
we're continuing to see clear examples where a model's harness is a major determinant of overall performance. with the same model, running on same task, it's easy to observe very different scores depe
A few days ago i wrote a post about how we use the LangGraph checkpointer to optimize storage. Today, LangGraph released v1.2 with a really nice feature: DeltaChannel, a new channel type that stores o
Agent observability enables monitoring of agent behavior, but to effectively improve agents through feedback loops, developers must actively collect observability data and generate synthetic examples
Come and join us, we're hiring across almost all roles 🚀https://www.langchain.com/careers Companies I'd consider going to if I ever had the urge to try something new. 1 Thinking Machines Lab 2 OpenAI
everyone would have a deeper appreciation for Agent Products that rock because of great Context/Harness Engineering if they… talked to: - LLM Base models - Post-Trained models with no harness (no tool
Great new drop by @LangChain’s @hwchase17 “If you do not know what the agent saw, what it did, and what happened next, you cannot reliably know what to improve” Traces give agents this insight and tha
I detected a bad Agent action, what do I do about it? this is pretty much the main question that will power the future’s Human+Agent driven improvement loops Gather data -> Mine Errors -> Find out whi
if you haven't read this one by @Vtrivedy10, it's a must read! great overview of what components a harness needs to support an agent for long running, long context tasks https://www.langchain.com/blog
Large language models are becoming increasingly costly to develop and operate, creating a need for open-source alternatives that reduce dependency on expensive proprietary models and make AI more acce
New from the LangSmith Signal: The vector space rapidly diversified since December. ✅ Managed cloud surged. MongoDB Atlas, Pinecone, + Qdrant tripled their combined share to 21%. ✅ Only 3 vector store
Not quietly. Loudly! 📣📣📣 deepagents-cli🔥 deepagents-cli is quietly becoming the best place to start coding with open weight models. we've been investing heavily in making it a harness that's truly mod
Observability helps power the agent improvement loop But it's not just observability! It's also feedback! You should be trying to get as much feedback (direct, indirect, generated) into your agent obs
Open models should compete on cost and specialization, not frontier benchmarks @natolambert puts it well: the right benchmark is savings in compute and time, especially for repetitive agent tasks deep
open-weight LLMs have come a long way on agent tasks! but the harness you wrap them in matters just as much as the model itself, and arguably the interface you use to drive that harness matters even m
.@Redisinc is returning to Interrupt! Say hello to their team at the expo hall to learn about their fast memory layer for chatbots and AI agents as well as their ready-to-use tools for building AI app
so well explained ! this is like applying Richard Sutton’s Bitter Lesson to agent systems. leverage comes from feedback loops, not handcrafted prompts. attaching feedback to traces is the missing piec
There’s a bunch of conflicting stances I don’t fully understand in the debate of Proprietary RL’d vs Open Harness, Model intelligence, and Agent Labs building harnesses for bespoke tasks Not all of th
Agent observability is enhanced by storing feedback alongside execution traces, transforming them from simple logs into a learning system that enables continuous improvement. This approach allows deve
'Traces everywhere. Feedback loop? Nowhere' @hwchase17 This is the part that gets skipped. Traces everywhere. Feedback loop? Nowhere. Building agents for clients and the hardest sell is 'we need to lo
Want to run the same harness across multiple interfaces? Try ACP. Deep Agents ships with it out of the box. open-weight LLMs have come a long way on agent tasks! but the harness you wrap them in matte
We’re hiring! https://www.langchain.com/careers Companies I'd consider going to if I ever had the urge to try something new. 1 Thinking Machines Lab 2 OpenAI 3 Anthropic 4 Cursor 5 Applied Intuition 6
CopilotKit has secured $27 million in funding to develop an agentic frontend stack that integrates AI agents with human-facing user interfaces, reflecting the belief that all future UI will incorporat
This post likely discusses Harrison Chase's preferred language model or AI system for handling coding tasks, potentially comparing different models' capabilities for programming work. As the creator o
You need feedback to help power the data flywheel! This 'feedback' can come from many sources: - direct user feedback (thumbs up/down) - indirect user feedback (code suggestions accepted) - llm as jud
changing an agent's data model mid-deployment is one of those problems that breaks things silently the langgraph runtime (which powers deepagents) has built-in support for backward compatibility and s
deepagents-cli is quietly becoming the best place to start coding with open weight models. we've been investing heavily in making it a harness that's truly model-agnostic, without compromising perform
deepagents you can run with a 'virtual filesystem' lets do lots of great context engineering tricks, without requiring an actual sandbox environment! @hwchase17 Where there's struggle is all of these
fleet is built on top of deepagents, a model agnostic base harness! try building multi-model agents in fleet today, or build your own deep agent with the underlying sdk! Not every step in an agent wor
@garrytan frames personal ai as the counter to extractive institutions: freedom to write your prompt and own your data. @hwchase17 made the same point from the dev side last month: if you don't own yo
Great conversation on harnesses from @hwchase17 at #GoogleCloudNext Is agent harness engineering the new prompt engineering? @LangChain co-founder and CEO, @hwchase17, joined us at #GoogleCloudNext to
i'll be speaking in NYC this thursday on memory for agents! i'll be talking about filesystem backed memory, owning memory with an open harness, background memory consolidation, and more! https://luma.
LangChain, a framework for developing applications with large language models, is advancing its capabilities and features according to insights from Harrison Chase, one of its key figures. This post l
LangGraph is battle tested against production weirdness. For example you can update an agent while it is running without breaking active threads Deep Agents inherits this resilience because it runs on
langgraph is the runtime that powers langchain and deepagents! we've been cooking on some new features: 1. node level error handlers 2. static + dynamic node timeouts 3. delta (diff based) channels fo
No longer are Fleet agents constrained to a single model. With multi-model support, you can build more efficient agents at scale http://langchain.com/langsmith/fleet Not every step in an agent workflo
Not every step in an agent workflow needs the same model. Fleet now lets you customize which model each sub-agent uses, so you can route simple tasks to fast/cheap models and keep stronger models for
Open sources harnesses powered by open source models I hear this take (usually from the model labs), but that will just mean more people turn to open source models, no? They’re already cheaper, if the
open weight models are having a moment @masondrxy is making deepagents cli a coding harness that works amazing with open weight models what is missing? what should he add? try it out! deepagents-cli i
production agents need to feel responsive in real time. two primitives make that possible: streaming and concurrent request control. on streaming: streaming is a first-class primitive in LangSmith's a
shipped Promptloop 🌀 Claude Code but for your prompts. a CLI agent that runs the full prompt-eval loop in one chat session: create test cases → run across models → generate reports → propose prompt →
Start today → https://docs.langchain.com/oss/python/deepagents/cli/overview deepagents-cli is quietly becoming the best place to start coding with open weight models. we've been investing heavily in m
This is probably better messaging than just “own your harness.” Yes, open models will need custom harnesses, but that is a means to an end. The end is utilizing models without being handcuffed to Anth
This is why I keep going on about LangGraph 1.2 alpha is looking great: node-level error handlers, timeouts, improved performance, and stronger typing LangGraph is the runtime layer you want underneat
Use open models in Fleet! Not every step in an agent workflow needs the same model. Fleet now lets you customize which model each sub-agent uses, so you can route simple tasks to fast/cheap models and
we're on an Open Model mission to help builders create world class agents >20x cheaper than what they have today a couple things have become evident recently: 1. The age of the token subsidy is being
This post discusses self-review capabilities in AI systems, where an AI model can evaluate and refine its own outputs before presenting them to users. According to Florian Jue from Listen Labs, this f
Ha love to see this port of flue to python Harness engineering is a fun time! Need more people exploring! 🎉Introducing PyFlue: The Python-Native Agent Harness Framework.🧰 💡Flue for Python: Fred K. Sch
I hear this take (usually from the model labs), but that will just mean more people turn to open source models, no? They’re already cheaper, if they have similar performance (because they’re less coup
🎉Introducing PyFlue: The Python-Native Agent Harness Framework.🧰 💡Flue for Python: Fred K. Schott @FredKSchott CEO of HTML has launched Flue: The Agent Harness Framework for TypeScript. It brings prog
it's always funny to me when these bot posts say we *just* launched deepagents... like nah man we've been carving this stone for close to a year now! Long-running agents almost always fail because the