OpenWork for developers: extend an open-source stack, or define one in a repo
OpenWork for developers is a free, open-source desktop app built on OpenCode, with 50+ providers and skills you can share. Kortix is the open-source AI Operating System: your agents, their skills, your company memory, and every connector in one git repo you own.
The configuration an agent reads
Agents, skills, connectors and memory are files in the repo, and the CLI is how you ship them. Each panel is quoted from the Kortix docs or the Kortix repository.
# yaml-language-server: $schema=https://kortix.com/schema/kortix.v2.schema.jsonkortix_version: 2default_agent: kortixenv: optional: [STRIPE_API_KEY]agents: kortix: file: agents/kortix.md connectors: all secrets: all skills: all kortix_permissions: alltriggers: - slug: daily-digest type: cron cron: "0 0 9 * * 1-5" timezone: Europe/Berlin prompt: Summarize yesterday's commits and open change requests.$ curl -fsSL https://kortix.com/install | bash$ kortix init my-agent && cd my-agent$ kortix ship$ kortix sessions new --prompt "Add a skill that drafts the weekly status report"$ kortix cr lsWhat OpenWork gives a developer
OpenWork is a free, open-source desktop app for macOS, Windows and Linux where AI agents work on your own files. It is built on OpenCode, so it reaches any model OpenCode supports: 50+ providers, your own API keys, or local models you run yourself (OpenWork on GitHub). It ships as a desktop app rather than a CLI, and an OpenWork MCP server exposes your assigned skills and MCP connections to an agent client you already run, through four tools: search_capabilities, execute_capability, list_skills and get_skill.
Where a repo-defined system differs
Kortix is the open-source AI Operating System, and it runs on OpenCode too. The difference is where the configuration lives and what a session does with it. Kortix keeps the whole company in one git repo you own. kortix.yaml declares the machine image, the connectors and the triggers, and secret names and grants. An agent is agents/<name>.md (behavior) plus an agents.<name> block in kortix.yaml (governance). Skills live at skills/<name>/SKILL.md, and memory is plain files today. Because the configuration is a repo, it is versioned and diffable, so a change to an agent or a skill is reviewed like code.
What a session does with the repo
OpenCode is the harness for both tools. In Kortix, one isolated sandbox per session. Each session has its own isolated machine and branch: the sandbox clones the repo, the agent works on a new branch, commits, and opens a change request. Session work reaches main through a change request. Merge is default-deny for agents. An agent can edit its own configuration on its session branch and propose the change, and a person approves it. That review gate is what makes a repo-defined system safe to extend, and the OpenWork vs Kortix comparison measures the same difference from the team's side.
Extending the system
Extending OpenWork means packaging skills, MCP servers and plugins and sharing them so a teammate imports the same setup in one click. Extending Kortix means adding a file and opening a change request: a new skill under skills/, a new trigger under triggers:, or a new connector. Connectors reach 3,000+ apps in a click, plus MCP, OpenAPI, Postman, GraphQL and raw HTTP. Connector credentials are brokered server-side and never enter the machine. A developer who wants the platform on their own hardware can read the self-hosting guide to see what a Kortix instance installs and what stays on disk.
Where each one fits
OpenWork suits a developer who wants an agent beside local files from a desktop app, on a base they already know, with a one-click way to share a setup with a teammate. Kortix suits a developer who wants the company configuration to be a repo the whole team can read: agents, skills, memory, connectors and triggers as files, each session on its own cloud computer, and every change landing through a review. If the starting question is what the desktop app does at all, what OpenWork is covers it, and the Kortix blog comparison walks the same decision from the team's side.
Questions developers ask about the two stacks
Short answers, each sourced to the product it describes.
How does a developer get started with Kortix?
Three commands.
curl -fsSL https://kortix.com/install | bashinstalls the CLI,kortix initscaffolds a project withkortix.yamland your agents and skills, andkortix shippushes the repo and brings it live (Kortix CLI reference).Can I bring my own model keys?
Yes. Kortix is model-agnostic. Any model provider with your own keys. Or the ChatGPT plan you already pay for. OpenWork works the same way with any model OpenCode supports.
Is Kortix a different harness from OpenWork?
OpenCode is the harness for both tools. The difference is what surrounds it: in Kortix the agents, skills, memory and connectors are files in a git repo you own, and each session runs on its own cloud computer.
Can I self-host Kortix instead of using the cloud?
Yes, and self-host is free. Run it on Kortix Cloud, in your VPC, or on your own on-prem network. The self-hosting guide lists what a self-hosted instance puts on disk.
Run your agents from one repo you own
Self-host for free, or start on managed cloud with any model and your own keys.
Open source · Any model, your keys · Self-host, VPC, or on-prem