Open source AnythingLLM alternatives, and which one fits your team
AnythingLLM is a local-first app for chatting with your documents. A team that needs agents, shared memory and connectors in one repo it owns should look at Kortix, the open-source AI Operating System. Here are the open-source alternatives, compared.
The shortlist, side by side
Four ways to run AI on your own terms, compared on source, models, hosting, and where your setup lives.
Rival details are from each project's own website or repository, checked October 2026.
The project layout
project (git repo + kortix.yaml)└─ session ──> isolated sandbox on branch "<uuid>"└─ agent commits + pushes└─ change request ──> merge ──> default branchA Kortix project is a git repo plus kortix.yaml. One isolated sandbox per session. Each session has its own isolated machine and branch.
The shortlist, one by one
AnythingLLM is a local-first app for chatting with your documents. It runs on your own machine, ingests PDFs and other files into a workspace, and ships a no-code agent builder, MCP compatibility and multi-user support on its Docker deployment (AnythingLLM on GitHub). Its memory is scoped per workspace and its configuration lives in the app, so what one workspace learns does not become shared company knowledge. AnythingLLM fits the job of document chat and retrieval on one machine or one Docker host. It is a weaker fit when the job is company-wide agent work that has to land as reviewed changes.
Kortix is the open-source AI Operating System: your agents, their skills, your company memory, and every connector live in one git repo you own (Kortix on GitHub). One isolated sandbox per session, and each session has its own isolated machine and branch, so an agent can install, run and break anything it likes because only what it commits survives. Session work reaches main through a change request, and merge is default-deny for agents, so a person reads the diff before anything lands. Connectors reach 3,000+ apps in a click, plus MCP, OpenAPI, Postman, GraphQL and raw HTTP, and connector credentials are brokered server-side and never enter the machine. Run it on Kortix Cloud, in your VPC, or on your own on-prem network; self-host is free, and it takes any model provider with your own keys.
Open WebUI is a self-hosted AI platform built to run entirely offline that connects to Ollama or any OpenAI-compatible API (Open WebUI docs). It is deep at the chat layer: granular roles and groups, LDAP, single sign-on and SCIM 2.0 provisioning, local retrieval over nine vector databases, plus Notes and Channels beside the conversation. Its memory is personal, so facts stay with one person's chat history, and configuration lives in the deployment's database rather than files. Open Terminal gives an agent a terminal and filesystem inside the chat, while per-user isolated terminals need the enterprise tier.
LM Studio is a desktop application for running local large language models on macOS, Windows or Linux, with llama.cpp and MLX runtimes, an OpenAI-like local server and a chat interface that can attach documents (LM Studio docs). It is the lightest way to try a model on your own hardware. It is a personal tool rather than a platform for a team: there is no shared company memory and no review gate, and the app is closed-source, so you cannot read or fork it.
How to choose
Pick AnythingLLM when your job is document chat and retrieval, on your own machine or a Docker host, and you want to start in minutes. Pick Open WebUI when you want a private, model-agnostic chat surface with identity and permissions for a team. Pick LM Studio when you mainly want to run local models on one computer.
Pick Kortix when the work is bigger than a chat window: agents that run on their own cloud computers, a shared company memory, connectors scoped per tool call, and every change landing as a diff a person approves. If you already run AnythingLLM and want to see how the move looks, the comparison of Kortix and AnythingLLM covers what transfers and what is rebuilt (Kortix vs AnythingLLM). For the wider field of self-hosted workspaces, see the six-way comparison on the Kortix blog (self-hosted AI workspace comparison).
Frequently asked questions
It depends on the job. For agents that run on their own cloud computers, a shared company memory and a review gate before work lands, pick Kortix. For document chat and retrieval on one machine or a Docker host, pick AnythingLLM. For a private, model-agnostic chat surface with roles and permissions, pick Open WebUI. For local models on one desktop, pick LM Studio.
Yes. AnythingLLM is built in the open, with its full source public on GitHub, so you can read, fork and audit it (AnythingLLM on GitHub). The desktop app is single-user, while multi-user support runs on the Docker deployment. Its memory is scoped per workspace rather than shared across a company.
Yes. AnythingLLM, Open WebUI and LM Studio are free to download and run, and Kortix self-host is free. Each option still needs a model, so you pay model or compute costs unless you run local models on hardware you already own.
Yes. AnythingLLM runs as a desktop app or a Docker deployment, Open WebUI installs with Docker, Python or Kubernetes, and LM Studio is a desktop app. Kortix self-hosts from one Docker Compose stack and also runs in your VPC or on your own on-prem network.
LM Studio is built for it, running local models with llama.cpp and MLX on macOS, Windows or Linux. AnythingLLM runs locally and connects to local or cloud providers, and Open WebUI connects to Ollama or any OpenAI-compatible API. Kortix takes any model provider with your own keys.
Own the system your agents run on.
Start on Kortix Cloud, self-host for free, and point it at any model with your own keys.
Open source · Any model, your keys · Self-host, VPC, or on-prem