Ollama Code

An agentic coding panel for your local Ollama models — a Claude Code / Codex–style chat experience that runs entirely on your machine.

Under the hood it drives the open-source OpenCode agent (Apache/MIT) as a headless server pointed at your Ollama server. You get a real agent — file edits, shell tools, permissions, multi-step reasoning — with no cloud model and no API key.

Demo

Why

The official Claude Code and Codex VS Code extensions are not open source, so they can't be adapted to local models. The CLIs behind several agents are open, though — and OpenCode in particular ships a headless server with a built-in Ollama provider. This extension wraps that server in a native chat panel and fills its model picker with the models you actually have installed.

Features

• Chat panel in the Activity Bar / secondary side bar (and "Open in Editor Tab" for parallel conversations)
• Streaming responses with a Claude-style timeline — thinking, tool steps, answer
• Reasoning blocks for thinking-capable models — folded away with how long the model thought once a turn ends; Show reasoning in the behavior menu hides them entirely
• Agent tools — file reads/edits, shell, search — surfaced as a quiet timeline: one row per call, timestamped, expandable only when there's a payload (edits open a diff)
• MCP servers — extend the agent with Model Context Protocol tools; servers you already configured for Claude Code (.mcp.json) or VS Code (.vscode/mcp.json) are picked up automatically. Type /mcp to see their live status
• Permission prompts — Allow once / Allow always / Deny, inline; plus a permissions picker in the composer: Auto (the default — safe actions run on their own; outside-the-workspace access and .env reads ask first), Manual (every tool call asks), or Bypass (nothing asks — for workspaces you trust)
• Model manager — load / eject Ollama models from the composer, with loaded state, context size, and capability badges (👁 vision / 🔧 tools)
• Multi-server — register, switch, edit, and remove Ollama servers; offline mode with a connection banner
• Context meter with compaction indicator, a Thinking slider in the behavior menu, and a + menu carrying image attachments for vision models plus the open file as excludable context
• Session history — persistent, resumable, auto-named; delete one or clear all
• Auto-context — loads the selected model with an adequate num_ctx so OpenCode's large system prompt doesn't overflow Ollama's small default window

Requirements

• VS Code 1.104+
• Ollama installed and running (default http://127.0.0.1:11434) with at least one model pulled — e.g. ollama pull llama3.2 (use a tool-capable model for the agent)

OpenCode is bundled — the matching platform binary ships inside the extension, so there's nothing extra to install and it works offline. Power users can point at their own build with ollamaCode.opencodePath; an install on your PATH or in ~/.opencode/bin is preferred over the bundled copy if present.

Quick start

1. Start Ollama and pull a model: ollama pull llama3.2 (or qwen3, mistral-small, etc.).
2. Install this extension (or run it from source — see below).
3. Click the Ollama icon in the Activity Bar.
4. Pick a model, type a task, hit Enter.

Beta channel

New features ship to the Marketplace pre-release channel first (odd minor versions, e.g. 0.11.x; stable releases use even minors). To try betas, open the extension's Marketplace page in VS Code and click Switch to Pre-Release Version — VS Code updates you along the beta track and you can switch back any time with Switch to Release Version.

Settings

MCP servers

The agent can call tools from MCP (Model Context Protocol) servers — browser automation, databases, issue trackers, docs, and more. OpenCode runs the servers; this extension just gathers them from wherever you've configured them and hands them over.

Where servers come from

Servers are merged from these sources, in increasing precedence (a later source wins on a name collision):

If you already use MCP with Claude Code or VS Code Copilot, those servers work here with nothing to re-enter. Use ollamaCode.mcpServers to add a server just for Ollama Code, or to override a discovered one.

Setting up a .mcp.json (shareable, per project)

Create .mcp.json at your project root — the same file Claude Code uses, so it's safe to commit and share with your team:

{
  "mcpServers": {
    // local (stdio) server — runs a command, talks over stdin/stdout
    "playwright": {
      "command": "npx",
      "args": ["-y", "@playwright/mcp@latest"]
    },
    // local server with a working dir and env var
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "."],
      "env": { "LOG_LEVEL": "info" }
    },
    // remote (http/sse) server, with a token pulled from the environment
    "docs": {
      "type": "http",
      "url": "https://example.com/mcp",
      "headers": { "Authorization": "Bearer ${MY_TOKEN}" }
    },
    // defined but off — won't be started
    "staging": {
      "command": "npx",
      "args": ["-y", "some-mcp-server"],
      "enabled": false
    }
  }
}

A .vscode/mcp.json is identical except the top-level key is servers instead of mcpServers (VS Code's convention) — both are supported.

What's supported

• ${VAR} references in env values, headers, and url are resolved from the environment before the server launches — keep secrets in your environment, not in the file.
• Transports: local (stdio) and remote (http/sse). Both the Claude Code field shape (command + args) and the VS Code shape are accepted and normalized for you.

Checking status — the /mcp command

Type /mcp in the chat to list your configured servers and their live status:

• 🟢 connected — running and its tools are available
• 🟡 disabled — defined but "enabled": false
• 🔴 failed — couldn't start/connect; the reason is shown (a bad server never blocks the chat)

Each row shows the transport (local/remote) and the command or URL it was configured with.

Notes

• Applying changes. Edits to ollamaCode.mcpServers (or VS Code's mcp setting) restart the agent automatically. Edits to the .mcp.json / .vscode/mcp.json files apply on the next Ollama Code: Restart OpenCode Server (or a window reload).
• Mind the context window. Each MCP server adds its tool schemas to every request. Local models have far less context than cloud ones (OpenCode's own system prompt + built-in tools already use ~11k tokens), so enable only the servers you need and raise ollamaCode.minContextLength if tools start crowding out the conversation.
• npx/uvx on PATH. Local servers launched with npx/uvx need Node and those tools on PATH. The extension augments PATH with common install locations (Homebrew, ~/.local/bin, nvm/fnm, bun, cargo), but if a server shows as failed, check Ollama Code: Show Logs.

How it works

VS Code webview (chat UI)
        │  postMessage
        ▼
Extension host (bridge)
        │  HTTP + SSE  (raw fetch)
        ▼
opencode serve  ──native ollama provider──▶  Ollama (/api/chat, local model)
   (OLLAMA_HOST + OPENCODE_CONFIG_CONTENT injected at launch)

The extension enumerates your installed models with Ollama's REST API (/api/tags, /api/show, /api/ps), then augments OpenCode's built-in ollama provider with those models (capabilities, context limit, num_ctx) via the OPENCODE_CONFIG_CONTENT environment variable — nothing is written to your workspace or global config. The active server is passed through OLLAMA_HOST. Model load/eject uses /api/generate with keep_alive and options.num_ctx.

Develop from source

npm install
npm run bundle:opencode      # fetch the pinned OpenCode binary into bin/ for your platform
npm run compile              # type-check + bundle (extension + webview)
# then press F5 in VS Code to launch the Extension Development Host
npm run package:vsix:bundled # build a platform .vsix with the binary embedded

The OpenCode binary is fetched at build time (pinned by opencodeVersion in package.json) and is never committed — bin/ is git-ignored. Bump that field to upgrade the bundled OpenCode. F5 also resolves the binary from bin/, so run bundle:opencode once before launching the dev host.

License

MIT