Ollama Free Coder

A fully-local LLM coding assistant powered by Ollama. Nothing leaves your machine.

Install (one-liner, recommended)

The VS Code extension is published on the official Marketplace:

• VS Code UI: open Extensions (Cmd/Ctrl+Shift+X), search for Ollama Free Coder, click Install.
• CLI:

  code --install-extension DenRaskovalov.ollama-free-coder

• Direct URL: https://marketplace.visualstudio.com/items?itemName=DenRaskovalov.ollama-free-coder

You still need a running Ollama server and at least one pulled model. The Ubuntu / macOS / Windows installers in scripts/ (described below) automate that part for you, but with the Marketplace install you can also just brew install ollama / winget install Ollama.Ollama / etc. and pull a model manually.

This repository ships three sister frontends that share the same ideas and core Ollama plumbing:

Frontend	Source	Installers	Docs
VS Code	src/	scripts/install-ubuntu.sh, scripts/install-macos.sh, scripts/install-windows.ps1	this README, DOCUMENTATION.md, ARCHITECTURE.md
Terminal app	src/cli.ts, src/cliTools.ts	npm bin after compile/package	this README, DOCUMENTATION.md
Vim / Neovim	vim/	vim/scripts/install-ubuntu.sh, vim/scripts/install-macos.sh, vim/scripts/install-windows.ps1	vim/README.md + :help ollama-coder

All six installers share the same environment-variable contract: CHAT_MODEL, COMPLETION_MODEL, EXTRA_MODELS, OLLAMA_HOST, SKIP_OLLAMA, SKIP_PULL. They each ensure the Ollama server is running, pull the configured models (failures are fatal, never silent), and verify them in /api/tags before installing the plugin.

If you don’t set CHAT_MODEL / COMPLETION_MODEL, the installer picks sensible defaults based on detected RAM:

Total RAM	Tier	Chat model	Completion model
< 6 GB	tiny	llama3.2:3b	qwen2.5-coder:0.5b-base
< 12 GB	small	llama3.1:8b	qwen2.5-coder:1.5b-base
< 20 GB	medium	qwen2.5:14b	qwen2.5-coder:1.5b-base
< 40 GB	large	qwen2.5:32b	qwen2.5-coder:7b-base
≥ 40 GB	huge	llama3.3:70b	qwen2.5-coder:7b-base

Override at any time:
CHAT_MODEL=qwen2.5:14b ./scripts/install-ubuntu.sh. The detection logic lives in scripts/pick-models.sh and scripts/pick-models.ps1 (same tier table), and prints ==> Detected N GB RAM (tier: X) -> ... so the choice is visible.

The rest of this README is about the VS Code extension and the standalone terminal app.

---

Standalone terminal app

The compiled package also provides a cross-platform CLI agent for Ubuntu, macOS, and Windows. Start it from a project folder and it treats that folder like VS Code's first workspace folder:

npm run compile
node out/cli.js
# or one-shot:
node out/cli.js "Generate a new C++ solution of LeetCode problem 2222"

Install only the terminal app and expose the global ofc / ollama-free-coder commands:

# Ubuntu/Linux
./scripts/install-cli-ubuntu.sh

# macOS
./scripts/install-cli-macos.sh

# Windows PowerShell
pwsh -ExecutionPolicy Bypass -File .\scripts\install-cli-windows.ps1

Run it from the current folder without globally linking it:

./scripts/run-cli-ubuntu.sh "Generate a new C++ solution of LeetCode problem 2222"
./scripts/run-cli-macos.sh "Play Radio Tapok music"
pwsh -ExecutionPolicy Bypass -File .\scripts\run-cli-windows.ps1 "Play Radio Tapok music"

When installed as an npm package, the same app is exposed as ollama-free-coder and ofc:

ofc "Play Radio Tapok music"
ofc --model qwen2.5:7b --enable-run-command

The CLI reuses the existing Ollama client, LLM router, agent loop, problem reference detector, music parser, web search, repo-map extraction, and SEARCH/REPLACE editor. Its terminal-specific tool executor (src/cliTools.ts) implements the same safety contract as the VS Code extension: all file paths are sandboxed to the startup folder, file writes require a y/N confirmation, and shell commands are disabled unless --enable-run-command is passed (then confirmed per command). Music requests open the configured streaming service in the local browser/app.

Useful options:

Option	Description
--model MODEL	Chat/agent model (default llama3.1:8b)
--router-model MODEL	Small intent-router model (default qwen2.5-coder:1.5b-base)
--endpoint URL	Ollama endpoint (default $OLLAMA_HOST or http://localhost:11434)
--cwd DIR	Workspace folder instead of the current directory
--enable-run-command	Enable the run_command tool, still with confirmation
--music-service amazon|spotify|youtube|apple	Default music service

---

Ollama Free Coder for VS Code

Features

• Inline ghost-text completion as you type, using a FIM-capable model (defaults to qwen2.5-coder:1.5b-base). Debounced and cancellable.
• Command history — every prompt you send is persisted. Press ↑/↓ in the chat input to walk through past commands shell-style, click History for a list view, or hover any past user message in the chat log for one-click resend and edit.
• Keyboard-driven model picker — click the Model button (or focus it with Tab) and use ↑/↓ to cycle models, Home/End to jump to first/last, PageUp/PageDown to skip 5, Enter to select, Esc to cancel. Mouse still works. The textarea’s own ↑/↓ history walk is independent — the keys never collide because they live on different focus targets.
• Chat sidebar (Activity Bar → robot icon) with streaming responses, a one-click "include current file/selection" toggle, and @mentions:
  • @src/foo.ts — attach a workspace file's contents to your question.
  • @selection — attach the current editor selection (or whole active file).
• Show-on-screen vs. write-to-file routing — prompts like “show me a C++ Hello World”, “explain Vector class in C++”, “what is a class in Python” always stream into the chat panel. Prompts like “add Vector class implementation to test.cpp” or “write a new file with Python fizzbuzz” are auto-routed through agent mode and create the file (with a confirm dialog). You can override either way by toggling agent mode.
• Web search built in — three ways to use it:
  1. Type /search QUERY (or /web, /google) in the chat input to run a web search directly and render the top hits inline. No LLM involved.
  2. Prompts with a web-search intent ("google the latest TypeScript", "what's new in Rust 1.85", "search the web for free SVG icons") are auto-grounded: results are fetched and prepended as context before the LLM answers.
  3. The agent can call web_search itself in agent mode. A small Search: DuckDuckGo (or Google) label in the chat input shows which backend is active.
• Play music — say “Play Radio Tapok from Amazon Music” or just “Play Five Finger Death Punch” and the extension opens the right streaming service for that artist/track in your browser. The Ollama model only extracts the artist and service from your words; the plugin builds the URL and opens it locally (no API keys, nothing leaves your machine). A named service wins; otherwise the ollamaCoder.musicService default (Amazon Music) is used. You can also type /play QUERY (alias /music) for a keyless, LLM-free shortcut, or run Ollama Free Coder: Play Music… from the command palette.
• Apply code blocks — hover any code block in chat for one-click Insert at cursor, Replace selection, Save… (with diff preview if the target file exists), and Copy. If the assistant emits a fence like ```ts src/foo.ts, Save… pre-fills that path.
• Agent mode (checkbox in the chat input) lets the model call workspace tools to answer multi-step questions and apply edits:
  • read_file, list_files, search_text — read-only context gathering.
  • get_open_editors — see what's open and what's selected.
  • web_search — search the public web. Default backend is DuckDuckGo (free, no key, no signup). If you set ollamaCoder.googleApiKey and ollamaCoder.googleCseId, the tool switches to Google Custom Search JSON API (Google's free tier: 100 queries/day).
  • write_file — create / overwrite files (always shows a confirm dialog; "Show diff first" opens a side-by-side preview before applying). Uses Ollama's native tool calling — works well with llama3.1:8b, qwen2.5:7b, qwen2.5-coder:7b, and other tool-capable models.
• Code actions on selection (right-click → Ollama Free Coder):
  • Explain Selection
  • Refactor Selection (replaces selection)
  • Fix Selection (replaces selection)
  • Add Docstrings / Comments (replaces selection)
  • Generate Unit Tests (opens new editor)
  • Ask About Selection… (free-form question)
• Status bar model switcher — click the Ollama: <model> badge to swap between any model you've pulled (ollama pull ...).
• Zero runtime npm deps — uses Node's built-in http for the Ollama API.

Default keybindings

Action	Shortcut
Open chat	Ctrl+Alt+O
Explain selection	Ctrl+Alt+E
Refactor selection	Ctrl+Alt+R

Settings

Setting	Default	Description
ollamaCoder.endpoint	http://localhost:11434	Ollama server URL
ollamaCoder.chatModel	llama3.1:8b	Model for chat & code actions
ollamaCoder.completionModel	qwen2.5-coder:1.5b-base	Model for inline completion (FIM-capable recommended)
ollamaCoder.enableInlineCompletion	true	Toggle ghost-text completion
ollamaCoder.completionDebounceMs	250	Delay before a completion request
ollamaCoder.maxCompletionTokens	128	Max tokens per completion
ollamaCoder.temperature	0.2	Sampling temperature
ollamaCoder.contextWindowChars	4000	Max chars of file context sent to the model
ollamaCoder.agentMaxSteps	8	Max tool-calling steps per agent turn
ollamaCoder.searchBackend	duckduckgo	duckduckgo (free, no key) or google (requires the two keys below)
ollamaCoder.googleApiKey	""	Optional Google API key for Custom Search JSON API (free 100/day)
ollamaCoder.googleCseId	""	Optional Google Programmable Search Engine id
ollamaCoder.enableMusic	true	Allow “play music” requests to open a streaming service locally
ollamaCoder.musicService	amazon	Default service when none is named: amazon / spotify / youtube / apple

Quick install on Ubuntu (24.04 / 26.04)

git clone https://github.com/local/ollama-coder-vscode
cd ollama-coder-vscode
./scripts/install-ubuntu.sh

The script will:

1. Check for Node.js ≥ 18 (and install via apt if missing).
2. Install Ollama (via the official installer) and systemctl enable --now ollama.
3. Pull the two default models (llama3.1:8b, qwen2.5-coder:1.5b-base).
4. npm install, compile TypeScript, package a .vsix, and install it into the code CLI.

Useful environment variables:

CODE_BIN=codium ./scripts/install-ubuntu.sh   # install into VSCodium instead of VS Code
SKIP_OLLAMA=1   ./scripts/install-ubuntu.sh   # skip Ollama install
SKIP_PULL=1     ./scripts/install-ubuntu.sh   # skip pulling models
CHAT_MODEL=qwen2.5-coder:7b-instruct ./scripts/install-ubuntu.sh

Manual build

npm install
npm run compile
npx vsce package -o ollama-free-coder.vsix
code --install-extension ./ollama-free-coder.vsix --force

Development

npm install
npm run watch       # incremental compile
# In VS Code: F5 to launch an Extension Development Host.

Source layout:

src/
  extension.ts           # activation, commands, status bar
  ollama.ts              # streaming HTTP client for /api/generate, /api/chat, /api/tags
  completionProvider.ts  # InlineCompletionItemProvider with FIM (prompt + suffix)
  codeActions.ts         # Explain / Refactor / Fix / Docstrings / Tests / Ask
  chatView.ts            # Webview-based chat sidebar
scripts/
  install-ubuntu.sh      # one-shot build + install for Ubuntu

Model recommendations

Use case	Suggested model	ollama pull
Inline completion (fast, FIM)	qwen2.5-coder:1.5b-base	ollama pull qwen2.5-coder:1.5b-base
Inline completion (better, FIM)	qwen2.5-coder:7b-base	ollama pull qwen2.5-coder:7b-base
Chat / code actions (small)	llama3.1:8b	ollama pull llama3.1:8b
Chat / code actions (better)	qwen2.5-coder:7b-instruct	ollama pull qwen2.5-coder:7b-instruct

The inline completion provider sends both a prompt (prefix) and suffix to /api/generate, which Ollama forwards to the model's fill-in-the-middle template. Use a *-base coder model for best results — instruct/chat models tend to wrap output in prose.

License

Apache-2.0 (see LICENSE).