Analysis updated 2026-05-18
Work through a large multi file refactor across several sessions without losing context.
Review AI proposed diffs and terminal output before approving any file or git change.
Connect a local Ollama model or your own provider API key so code processing stays local.
Connect custom MCP servers so the agent can use your own tools with approval based checks.
| mereyabdenbekuly-ctrl/clodex-ide | portabase/portabase | skevy/wobble | |
|---|---|---|---|
| Stars | 849 | 843 | 861 |
| Language | TypeScript | TypeScript | TypeScript |
| Last pushed | — | — | 2023-02-10 |
| Maintenance | — | — | Dormant |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 3/5 | 3/5 | 2/5 |
| Audience | developer | ops devops | developer |
Figures from each repo's GitHub metadata at analysis time.
The current preview build is unsigned and not notarized, so the OS will show extra security warnings before it runs.
CLODEx is a desktop app for coding with AI help, built for programmers who work on long, multi step engineering tasks rather than quick one off questions. Instead of treating each request as a fresh chat that loses its context, CLODEx keeps a task's files, git history, terminal, browser view, and connected tools together in one workspace that you can leave and come back to later without starting over. The idea behind the project is that AI output should be treated as a suggestion, not something that runs unchecked. When you ask the built in agent to explain, plan, change, or review code, it can inspect your files and use approved local tools, but sensitive actions can be set to require your approval before they run. You get to see pending edits and line by line diffs, read terminal output, check the embedded browser, and decide whether to accept, change, or continue the work. It supports several ways of connecting to AI models: signing in with a CLODEx account, using your own API key from a provider, pointing it at any OpenAI compatible endpoint, or running local models through Ollama. It also supports the Model Context Protocol (MCP), letting you connect external tools, servers, and prompts, with the same approval based safety checks applied. The project is open source and local first, meaning your code and work stay on your machine rather than being routed entirely through a third party service. It is written mostly in TypeScript and packaged as a desktop app, and currently ships a Technical Preview build for macOS, Windows, and Linux, though that build is unsigned and not notarized, meaning your operating system may show extra security warnings before it will run. It is released under the AGPL 3.0 license, which requires that modified versions shared with others, including over a network, also be released as open source.
A local first desktop AI coding app that keeps files, git, terminal, browser, and tools together in one task workspace you can pause and resume.
Mainly TypeScript. The stack also includes TypeScript, Electron, Node.js.
Released under AGPL 3.0, meaning modified versions, including ones offered over a network, must also be shared as open source.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.