Agentic IDE

AI code editor with agentic execution

Coder is the IDE inside VIMS: Monaco editor, integrated terminal, Git, and an AI agent that reads your repo, edits files, runs tests, and ships — locally.

Inside VIMS Desktop · macOS · Windows · Linux · ChromeOS · Free

✓ tests passed

What it does

Coder puts the editor, terminal, Git, project search, and the agent loop in one workspace. Ask for an outcome; the agent explores the codebase, makes changes, runs commands, and reports back — with every tool call visible and approvable.

01
Monaco editorThe VS Code editor engine with syntax, diagnostics, and multi-file editing.
02
Integrated terminal and GitRun builds and tests, stage and commit, without leaving the window.
03
Repo-aware agentReads the project, searches code, edits files, and executes multi-step tasks.
04
MCP toolsAny connected MCP server (GitHub, Postgres, Playwright, …) is callable from the agent.
05
Skills and rulesProject SKILL.md workflows and rules shape how the agent works in each repo.
06
Local models welcomeRoute to Ollama for fully offline coding, or to hosted models with your keys.

How it works

  1. Open a folder or clone a repo
  2. Describe the change you want
  3. Review the diff, run tests, commit

FAQ

Does the agent run commands without asking?

Consequential actions go through tool approvals you configure per project. Read-only exploration runs freely.

Which models can it use?

Any provider configured in Settings — local Ollama models, OpenRouter, or your own API keys.

How does VIMS Coder connect to the rest of VIMS?

It shares your identity from keys.vims.com and is reachable by agents through MCP and Flow, so it works alongside every other VIMS app.

Related

AI code editor with agentic execution

Coder puts the editor, terminal, Git, project search, and the agent loop in one workspace. Ask for an outcome; the agent explores the codebase, makes changes, runs commands, and reports back — with every tool call visible and approvable.