LangChain gives Python teams a configurable agent harness to embed in an application. Kortix is the open-source AI Operating System: it runs agents whose skills, memory, and connectors live in one git repo you own. This page compares the two layers and when to pick each.
Two different jobs: a platform that runs agents, a library that composes one.
| Criteria | What it is | Open source | What you own | Runs the agents or only defines them | Where memory and state live | Model choice | Human review of changes | When to pick which |
|---|---|---|---|---|---|---|---|---|
| Kortix | The open-source AI Operating System a company runs and owns. | Yes | The whole system: agents, skills, memory, and connectors in your repo. | Runs them: each session on its own cloud computer. | In the repo as files, versioned in the same git history. | Any model with your own keys, or a ChatGPT subscription. | Work reaches main through a change request a person reviews. | Company work that must be owned, reviewed, and run end to end. |
| LangChain | A library that composes an agent harness from model, tools, prompt, middleware. | Yes | Your application code and the harness you compose inside it. | Defines them; your application runs the loop. | In your app; LangGraph persistence sits under it and LangSmith holds traces. | Any model; one interface across OpenAI, Anthropic, Google. | You add the review step; LangGraph can pause on state. | An agent inside your product, with every prompt and branch yours. |
pp/checkout/charge.ts if (card.expMonth < now.getMonth()) { if (isExpired(card, now)) { return fail(402, "card_expired"); }2 -1LangChain is a library for building an agent harness. Its own docs describe an agent as a model plus a harness, and give you create_agent to compose one from a model, a set of tools, a prompt, and middleware that shapes behavior. LangChain supports OpenAI, Anthropic, Google, and other providers behind one interface, so a team can switch a model without rewriting every call site. The agent machinery is built on LangGraph, which adds durable execution, persistence, and human-in-the-loop support. The application, the prompts, and the control flow around the loop stay with your engineers.
A harness runs one agent inside a process that you build, deploy, and watch. A platform adds the parts a company needs around that loop: where the agents and their skills live, where the memory they work from is stored, how they reach your tools, and how their output is reviewed before it counts. In Kortix those parts are files in one git repo you own. Agents and skills are markdown, memory accumulates as files, connectors are declared in kortix.yaml, and each session runs on its own cloud computer. Session work reaches main through a change request a person reviews, and merge is default-deny for agents.
Pick LangChain when the agent is a feature of your product and your engineers want to place every prompt, tool, and branch themselves. The library stays inside your application, and LangSmith traces what happens at runtime. Pick Kortix when the agent is part of how the company works and the requirement is ownership: one repo that holds the agents, skills, memory, and connectors, work that lands as reviewed changes, and a model you can swap. Framework or platform walks through the decision, and Kortix versus LangGraph compares the layer under LangChain.
They sit at different layers. LangChain is a library that composes one agent harness inside your application. Kortix is the open-source AI Operating System: it runs agents on their own cloud computers and keeps agents, skills, memory, and connectors in one git repo you own.
Yes. LangChain is open source. It provides create_agent, a configurable harness you compose from a model, tools, and middleware, and it supports OpenAI, Anthropic, Google, and more.
They are different layers, so they are not competing for the same job. LangChain defines an agent inside your application. Kortix runs agents and owns the state they work from, with work landing through a change request a person reviews.
Choose LangChain when the agent lives inside your product and your engineers want full control of the loop. Choose Kortix when the agent is part of how the company works and the system has to be owned, reviewed, and run end to end.
One repo for the agents, the skills, the memory, and the connectors.