lollama-ai-assistant

# LoLLaMA - Open Source AI Browser Assistant

Your personal AI browser assistant. Chat with local or cloud LLMs, modify websites, and create widgets — all from a sidebar. Privacy-first, no account required.

[![Chrome Web Store](https://img.shields.io/chrome-web-store/v/nkbpbdfjaaeopaooikbodchghaeipfaa?label=Chrome%20Web%20Store)](https://chromewebstore.google.com/detail/lollama-personal-ai-brows/nkbpbdfjaaeopaooikbodchghaeipfaa)
[![Chrome Web Store Users](https://img.shields.io/chrome-web-store/users/nkbpbdfjaaeopaooikbodchghaeipfaa)](https://chromewebstore.google.com/detail/lollama-personal-ai-brows/nkbpbdfjaaeopaooikbodchghaeipfaa)
[![Chrome Web Store Rating](https://img.shields.io/chrome-web-store/rating/nkbpbdfjaaeopaooikbodchghaeipfaa)](https://chromewebstore.google.com/detail/lollama-personal-ai-brows/nkbpbdfjaaeopaooikbodchghaeipfaa)
![License](https://img.shields.io/badge/License-MIT-yellow)

## What It Does

**LoLLaMA** is a Chrome extension that puts an AI assistant in your browser sidebar. It connects to **local LLMs** (Ollama, LM Studio, etc.) or **cloud APIs** (OpenAI, Anthropic, Google, etc.) and lets you:

- **Chat** — Ask questions, get explanations, with full page context
- **Modify** — Change any website with natural language ("hide YouTube Shorts", "add dark mode")
- **Create** — Build interactive widgets (timers, calculators, notes) that overlay any page

All data stays local. Your API keys are stored in your browser and never leave your device.

## Install

### From Chrome Web Store

[**Install LoLLaMA**](https://chromewebstore.google.com/detail/lollama-personal-ai-brows/nkbpbdfjaaeopaooikbodchghaeipfaa) — one click, no setup.

### From source

```bash
git clone https://github.com/anugotta/lollama-ai-assistant.git
cd lollama
```

1. Open Chrome → `chrome://extensions/`
2. Enable **Developer mode** (top right)
3. Click **Load unpacked** → select the cloned folder
4. Click the LoLLaMA icon in the toolbar to open the sidebar

### Configure your AI

Open Settings (gear icon) and pick one:

| Option | What you need |
|--------|--------------|
| **Local LLM** | A running server — [Ollama](https://ollama.com), [LM Studio](https://lmstudio.ai), llamafile, vLLM, etc. |
| **Cloud API (BYOK)** | Your own API key from any supported provider |
| **Custom endpoint** | Any OpenAI-compatible API URL |

Click **Test Connection** → **Save Settings** → start chatting.

## Supported Providers

### Local LLMs (100% private, no API key needed)

| Provider | Port | Install |
|----------|------|---------|
| [LM Studio](https://lmstudio.ai) | `:1234` | Download app → Developer tab → Start Server |
| [Ollama](https://ollama.com) | `:11434` | `ollama serve && ollama pull llama3.3` |
| [Jan.ai](https://jan.ai) | `:1337` | Enable API server in settings |
| [LocalAI](https://localai.io) | `:8080` | `docker run -p 8080:8080 localai/localai` |
| [llamafile](https://github.com/Mozilla-Ocho/llamafile) | `:8080` | Run the `.llamafile` binary |
| [vLLM](https://docs.vllm.ai) | `:8000` | `vllm serve <model> --port 8000` |
| [Text Generation WebUI](https://github.com/oobabooga/text-generation-webui) | `:5000` | Start with `--api` flag |
| [KoboldCpp](https://github.com/LostRuins/koboldcpp) | `:5001` | Run with `--api` flag |

### Cloud APIs (Bring Your Own Key)

| Provider | Suggested models | Pricing |
|----------|-----------------|---------|
| [OpenAI](https://platform.openai.com/api-keys) | GPT-4o, o3-mini, o1 | Pay per token |
| [Anthropic](https://console.anthropic.com/settings/keys) | Claude Sonnet 4, Opus 4 | Pay per token |
| [Google](https://aistudio.google.com/apikey) | Gemini 2.0 Flash, 1.5 Pro | Free tier + paid |
| [xAI](https://console.x.ai/) | Grok 3, Grok 3 Mini | Pay per token |
| [Groq](https://console.groq.com/keys) | Llama 3.3 70B, DeepSeek R1 | Free tier |
| [DeepSeek](https://platform.deepseek.com/api_keys) | DeepSeek V3, R1 | Very affordable |
| [Mistral](https://console.mistral.ai/api-keys/) | Mistral Large, Codestral | Pay per token |
| [OpenRouter](https://openrouter.ai/keys) | All major models | Aggregated pricing |
| [Together AI](https://api.together.xyz/settings/api-keys) | Llama, Qwen, DeepSeek | $25 free credit |
| [Fireworks AI](https://fireworks.ai/api-keys) | Llama, Qwen, DeepSeek | Pay per token |
| [Perplexity](https://www.perplexity.ai/settings/api) | Sonar Pro, Sonar Reasoning | Pay per token |
| [Cohere](https://dashboard.cohere.com/api-keys) | Command R+ | Free tier + paid |
| [SambaNova](https://cloud.sambanova.ai/apis) | Llama 3.3 70B | Free tier |
| [NanoGPT](https://nano-gpt.com/api) | GPT-4o, Claude, Grok | Pay-as-you-go |

> Model lists are **suggestions only** — you can type any model name your provider supports.

## Features

### Modify websites with natural language

```
"Hide YouTube Shorts on this page"
"Make this page dark mode"
"Remove the sidebar on Reddit"
"Make all text larger and high-contrast"
```

Changes are applied via CSS/JS injection. You can save modifications per domain so they auto-apply on future visits.

### Create widgets (Doodles)

```
"Build a Pomodoro timer"
"Create a color picker tool"
"Build a reading progress bar"
"Create a sticky note widget"
```

Widgets are self-contained HTML/CSS/JS overlays. Drag to move, save to your library, reuse on any page.

### Context-aware chat

The AI sees the current page context (URL, elements, selected text) so it can give relevant answers without you copy-pasting.

### Text tools

Select any text on a page → use the tooltip or right-click menu to explain, fix grammar, summarize, translate, or ask the AI about it.

### Chat history

Conversations are saved as sessions. Open the History panel (clock icon) to load, continue, or delete past chats. Each sidebar open starts fresh — history is there when you need it.

### Keyboard shortcuts

| Shortcut | Action |
|----------|--------|
| `Ctrl+Shift+L` | Toggle sidebar |
| `Ctrl+Shift+N` | New chat |
| `Ctrl+Shift+D` | Toggle dark mode on page |

## Project Structure

```
lollama/
├── manifest.json              # Chrome extension manifest (MV3)
├── src/
│   ├── background/            # Service worker
│   │   └── service-worker.js
│   ├── sidebar/               # Main chat UI
│   │   ├── sidebar.html
│   │   ├── sidebar.css
│   │   └── sidebar.js
│   ├── options/               # Settings page
│   │   ├── options.html
│   │   ├── options.css
│   │   └── options.js
│   ├── content/               # Content scripts (page injection)
│   │   ├── content-script.js
│   │   ├── content-styles.css
│   │   └── doodle-frame.html
│   ├── shared/                # Shared modules
│   │   ├── constants.js       # Provider configs, system prompts
│   │   ├── storage.js         # Chrome storage helpers
│   │   └── ai-service.js      # API call layer
│   └── assets/                # Icons and images
├── PRIVACY_POLICY.md
├── TERMS_OF_SERVICE.md
├── CONTRIBUTING.md
└── LICENSE
```

## Development

### Prerequisites

- Chrome (or any Chromium-based browser)
- A local LLM server **or** a cloud API key for testing

### Local development

1. Clone the repo and load it as an unpacked extension (see [Quick Start](#quick-start))
2. Make changes to files in `src/`
3. Go to `chrome://extensions/` → click the **reload** button on the LoLLaMA card
4. Test your changes

No build step required — it's plain HTML/CSS/JS with ES modules.

### Adding a new cloud provider

Edit `src/shared/constants.js`:

```javascript
export const CLOUD_PROVIDERS = {
  // ...existing providers...
  yourprovider: {
    name: 'Your Provider',
    url: 'https://api.yourprovider.com/v1',
    models: [
      { id: 'model-id', name: 'Model Name', description: 'Short description' }
    ],
    keyPlaceholder: 'your-api-key',
    keyUrl: 'https://yourprovider.com/api-keys',
    pricing: 'Pricing info'
  }
};
```

Then add an `<option>` in `src/options/options.html` under the `#cloud-provider` select. That's it — the model field is free-text so users can type any model name.

#

[truncated…]