SymGraph

On this page

Getting Started with BinAssist

This guide will help you install BinAssist, configure an LLM provider, and run your first analysis.

Prerequisites

Before installing BinAssist, ensure you have:

Installation

Step 0: Running Windows? Read this first

BinAssist works on Windows, but the MCP SDK has some dependencies that require manual effort to install properly. Please refer to: BinAssist on Windows

Step 1: Install the Plugin

BinAssist can be installed from the Binary Ninja Plugin Manager or manually:

Option A: Plugin Manager (Recommended)

  1. Open Binary Ninja
  2. Go to Edit > Preferences > Plugin Manager
  3. Search for “BinAssist”
  4. Click Install
  5. Restart Binary Ninja

Option B: Manual Installation

  1. Download or clone the BinAssist repository
  2. Copy the BinAssist folder to your Binary Ninja plugins directory:
    • Linux: ~/.binaryninja/plugins/
    • macOS: ~/Library/Application Support/Binary Ninja/plugins/
    • Windows: %APPDATA%\Binary Ninja\plugins\
  3. Restart Binary Ninja

Step 2: Install Dependencies

Open a terminal in the BinAssist plugin directory and run:

pip install -r requirements.txt

This installs the required Python packages: - openai - OpenAI and compatible API client - anthropic - Anthropic Claude API client - httpx - HTTP client for API calls - mcp - Model Context Protocol client - whoosh - Full-text search for RAG

Initial Configuration

After installation, you need to configure at least one LLM provider.

Opening BinAssist

  1. Load any binary in Binary Ninja
  2. Look for the robot icon in the sidebar (left side - below the Sidekick icon)
  3. Click the icon to open the BinAssist panel

BinAssist Sidebar Icon

Accessing Settings

  1. In the BinAssist panel, click the Settings tab (last tab)
  2. You’ll see the LLM Providers section at the top

Setting Up an LLM Provider

BinAssist supports multiple LLM providers. Choose the one that best fits your needs:

Option 1: Ollama (Local, Free, Private)

Ollama runs models locally on your machine, ensuring privacy and avoiding API costs.

Step 1: Install Ollama

# Linux/macOS
curl -fsSL https://ollama.ai/install.sh | sh

# Windows: Download from https://ollama.ai/download

Step 2: Pull a Model

# General purpose model
ollama pull llama3.1:8b

# Or a reasoning model (recommended for complex analysis)
ollama pull gpt-oss:20b

# Start the server
ollama serve

Step 3: Configure in BinAssist

  1. In the Settings tab, click Add in the LLM Providers section
  2. Fill in the fields:
    • Name: Ollama Local
    • Type: Select Ollama
    • Model: llama3.1:8b (or your chosen model)
    • URL: http://localhost:11434
    • API Key: Leave empty
    • Max Tokens: 8192
  3. Click Save
  4. Click Test to verify the connection

Option 2: OpenAI Platform API

Use OpenAI’s models with a paid API key.

Step 1: Get an API Key

  1. Go to platform.openai.com
  2. Sign up or log in
  3. Navigate to API Keys
  4. Create a new API key
  5. Copy the key (you won’t be able to see it again)

Step 2: Configure in BinAssist

  1. Click Add in the LLM Providers section
  2. Fill in the fields:
    • Name: OpenAI
    • Type: Select OpenAI Platform API
    • Model: gpt-5.2-codex
    • URL: Leave empty (uses default)
    • API Key: Paste your API key
    • Max Tokens: 20000
  3. Click Save
  4. Click Test to verify

Option 3: Anthropic Platform API

Use Anthropic’s Claude models with a paid API key.

Step 1: Get an API Key

  1. Go to console.anthropic.com
  2. Sign up or log in
  3. Navigate to API Keys
  4. Create a new API key
  5. Copy the key

Step 2: Configure in BinAssist

  1. Click Add in the LLM Providers section
  2. Fill in the fields:
    • Name: Anthropic Claude
    • Type: Select Anthropic Platform API
    • Model: claude-sonnet-4-5 (or claude-opus-4-5)
    • URL: Leave empty (uses default)
    • API Key: Paste your API key
    • Max Tokens: 20000
  3. Click Save
  4. Click Test to verify

Option 4: OAuth Providers (Claude Pro/Max or ChatGPT Pro/Plus)

If you have a Claude Pro/Max or ChatGPT Pro/Plus subscription, you can use OAuth authentication instead of an API key.

For Claude Pro/Max:

  1. Click Add in the LLM Providers section
  2. Select Type: Anthropic OAuth
  3. Fill in a Name and Model (e.g., claude-sonnet-4-5)
  4. Click Authenticate
  5. A browser window will open for you to sign in to your Anthropic account
  6. After authorization, the credentials will be saved automatically
  7. Click Save

For ChatGPT Pro/Plus:

  1. Click Add in the LLM Providers section
  2. Select Type: OpenAI OAuth
  3. Fill in a Name and Model (e.g., gpt-5.2-codex)
  4. Click Authenticate
  5. A browser window will open for you to sign in to your OpenAI account
  6. After authorization, the credentials will be saved automatically
  7. Click Save

Setting the Active Provider

After adding providers, select the one you want to use:

  1. In the Active Provider dropdown at the bottom of the LLM Providers section
  2. Select your configured provider
  3. This provider will be used for all BinAssist operations

Your First Analysis

Now that BinAssist is configured, let’s run your first analysis.

Step 1: Load a Binary

  1. Open a binary file in Binary Ninja (File > Open)
  2. Wait for the initial analysis to complete

Step 2: Navigate to a Function

  1. In the Functions list (left panel), click on any function
  2. Or use Go to Address (G key) to jump to a specific location

Step 3: Explain the Function

  1. Open the BinAssist sidebar (robot icon)
  2. Click the Explain tab
  3. Click the Explain Function button
  4. Wait for the LLM response to stream in

You should see: - A detailed explanation of what the function does - Security analysis panel with risk assessment - Activity profile and detected API patterns

Step 4: Ask a Question

  1. Switch to the Query tab
  2. Type a question in the input field, for example:
    • “What does this function do?”
    • “Are there any security concerns here?”
    • “What functions does this call?”
  3. Press Enter or click Send
  4. Watch the response stream in

Next Steps

Now that you have BinAssist working, explore these features:

Troubleshooting

“Connection failed” when testing provider

No response from LLM

Plugin not appearing in sidebar

Slow responses

For additional help, check the GitHub Issues page.