Skip to content

Best Self-Hosted AI Coding Agents Tools

Self-hosted AI coding tools put the interface and the data on your server while you bring the model. The five picks cover different jobs. OpenHands is a browser control centre that runs its own agent or third-party ones such as Claude Code and Codex, installed with npx and wanting about 4 GB of RAM. OpenCode is an MIT terminal-first agent whose opencode web mode serves the same sessions in a browser behind basic auth, and our snippet runs its pinned official image. goose is a general-purpose agent written in Rust and licensed Apache-2.0, with a command-line tool, a desktop app and more than 15 model providers; it has no web interface of its own, and our snippet runs the CLI from the project's published image. bolt.diy generates web apps from a prompt and runs them in a WebContainer in your browser; it builds from a git tag, and its Dockerfile gives the build a 4 GB Node heap. PR-Agent is not a coding agent: it reviews and describes merge requests from a webhook on GitHub, GitLab, Bitbucket, Azure DevOps or Gitea. Four are MIT-licensed and goose is Apache-2.0, none ships a model, and every one needs an LLM API key or a local OpenAI-compatible server before it does anything. Treat the agents as remote shells: OpenHands, OpenCode and goose can run commands, so firewall them, set their passwords where they have one, and give them a project directory rather than your whole VM. PR-Agent sends diffs to the model provider you choose.

We earn commissions when you shop through the links below. Full disclosure →

AI Coding Agents apps

AI Coding Agents

bolt.diy

An open-source, self-hosted fork of StackBlitz's Bolt.new: describe an app in a chat, pick a model from more than 20 providers (OpenAI, Anthropic, Google, OpenRouter, Groq, Ollama and others) and bolt.diy writes the files, installs packages and runs the project in a WebContainer inside your browser. The server only relays model calls; the code runs client-side, and you bring your own API keys.
TypeScriptReactVite
easyRead guide →
AI Coding Agents

goose

A general-purpose AI agent written in Rust that runs on your own machine, with a desktop app, a command-line interface and an API. It connects to more than 15 model providers (Anthropic, OpenAI, Google, Ollama, OpenRouter, Azure, Bedrock and others) and to extensions through the Model Context Protocol. It has no web interface of its own; `goose serve` exposes an ACP server over HTTP and WebSocket. The CLI is a tool you run in a terminal, and the container image the project publishes is that CLI with git and curl added, so it needs a model provider configured before it can do any work.
Rust
easyRead guide →
AI Coding Agents

OpenCode

An open-source AI coding agent that runs in a terminal, a desktop app or, with opencode web, in your browser. It supports 75 or more LLM providers through the AI SDK and Models.dev, including local models, with sessions you can resume. The web server can be protected with HTTP basic auth, and a terminal can attach to the same sessions with opencode attach.
TypeScript
easyRead guide →
AI Coding Agents

OpenHands

A self-hosted control center for AI coding agents: describe a task in natural language and OpenHands writes code, runs commands, and iterates on its own. It ships as Agent Canvas, an npx-installable web UI that runs its own agent or third-party agents (Claude Code, Codex, and other Agent-Client-Protocol-compatible agents), and you bring your own LLM API key (Anthropic, OpenAI, or a self-hosted OpenAI-compatible endpoint).
TypeScriptReactPython
moderateRead guide →
AI Coding Agents

PR-Agent

An open-source AI pull-request reviewer, the project formerly published as Qodo's (CodiumAI's) PR-Agent and now under The-PR-Agent organisation. Run as a webhook server, it listens to GitLab merge-request events and answers slash commands in comments such as /describe, /review, /improve and /ask, posting an AI-written description, review and code suggestions on the merge request. It works with GitHub, GitLab, Bitbucket, Azure DevOps and Gitea through separate server targets, and calls an LLM you configure (OpenAI, Anthropic, Gemini, Ollama and others via LiteLLM).
PythonFastAPI
moderateRead guide →
Where to host itaffiliate disclosure
Kamaterafree trial
From $4/mo · 1 vCPU / 1 GB / 20 GB · US + EU + Asia
Start free on Kamatera → (opens in new tab)
DigitalOceanalso works on
From $6/mo · 1 vCPU / 1 GB / 25 GB · US + EU + Asia
Deploy on DigitalOcean → (opens in new tab)

Paid link — we earn a commission if you shop through it.

Common questions

How much RAM do self-hosted apps in AI Coding Agents need?

Across the 5 apps in AI Coding Agents, minimums run from 256 MB (goose) to 4096 MB (bolt.diy).

Which app in AI Coding Agents uses the least RAM?

Of the 3 apps in AI Coding Agents we measured, OpenCode used the least: 182 MB at idle (first boot, measured Nov 2026).

Do self-hosted apps in AI Coding Agents support Docker?

4 of the 5 apps in AI Coding Agents list Docker support.

Analytics cookies only (Google Analytics, PostHog). No ad networks, no cross-site tracking. See our privacy policy.