Skip to content

Getting Started

Follow these three steps to get SnapPoint up and running in under a few minutes.

Step 1: Installation

SnapPoint is available for Chrome and Edge. Choose your store:

Step 2: Model Setup (REQUIRED)

SnapPoint is an AI-powered "shell"; you need to connect an AI "brain" before it can generate summaries.

SnapPoint Model Library

Local Models

  1. Runner Setup: Download and install a local LLM runner such as llama.cpp, or LM Studio or Ollama.
  2. Download Model: Choose a model (e.g., Qwen, Gemma) and download a version compatible with your hardware.
  3. Start Local Server: Ensure the runner's local API server is active.
  4. Connect: Open SnapPoint Settings → Model Settings:
    • Base URL: Enter your runner's API endpoint (e.g., http://localhost:1234 for LM Studio or http://localhost:11434 for Ollama).
    • Model ID: Enter the exact id of the model you configured in your model server.

TIP

Quick Setup: You can also use the Presets dropdown in Model Settings to automatically fill in the correct Base URL and configuration for providers like LM Studio, Ollama, vLLM, or cloud providers like DeepSeek and ZhipuAI.

API Providers

If you prefer a hosted model, enter your API key in Settings → Model Settings.

TIP

Quick Setup: Similar to local models, you can use the Presets dropdown to automatically configure endpoints and recommended model names for common providers like OpenAI (GPT-4o, o3-mini), Groq, or Kimi.

Action Required

The summarization feature will not work until a model is successfully configured and tested.

Step 3: Your First Summary (Quick Start)

Now that your model is connected, you're ready to summarize!

SnapPoint in Action
  1. Highlight text: Move your mouse over any content on a webpage.
  2. ✨ Snap it: Click the Snap Button that appears next to your hovered content.
  3. Read: The Side Panel will open and stream your summary in real-time.

INFO

You can also click the expand arrow on the Snap Button to widen the selection to the parent DOM block before summarizing, or use the ▼ agent picker to choose a specific sub-agent before you snap.

Further Reading