Eigent (Desktop Agent)

Route your Eigent desktop agent through LLM Router to automatically prune context, redact local secrets, and leverage model fallbacks.

---

Documentation Index

Fetch the complete documentation index at: https://docs.llmrouter.app/llms.txt Use this file to discover all available pages before exploring further.

Eigent & LLM Router

Eigent is an open-source “coworker” agent that runs directly on your desktop. It uses a multi-agent architecture to perform browser automation, terminal automation, and file system operations, acting much like a human worker operating in a real desktop environment.

Because Eigent interacts heavily with your local machine, routing its traffic through LLM Router is highly recommended. It ensures that any local .env files or API keys it accidentally reads are redacted before reaching external LLM providers, and it compresses bloated terminal/browser outputs to save you massive amounts of input tokens.

Important: When using LLM Router with Eigent, you must configure the Custom API Host to point to https://api.llmrouter.app/v1 to ensure API compatibility.

Step 1: Installing Eigent

Choose the installation method that best suits your needs:

Official Download (Recommended)

1. Download Eigent

Visit eigent.ai and download the latest version for your platform (macOS 11+ or Windows).

2. Install the Application

• macOS: Open the downloaded .dmg file and drag Eigent into your Applications folder.
• Windows: Run the downloaded .exe installer and follow the on-screen instructions.

3. Launch Eigent

Open the application to get started.

Run from Source

For developers who want to run Eigent locally from source code.

Prerequisites:

• Node.js >= 18.0.0
• Python >= 3.12
• Docker (recommended) or PostgreSQL 15

1. Clone the Repository

git clone https://github.com/eigent-ai/eigent.git
cd eigent

2. Start Backend Services

cd server
cp .env.example .env
docker compose up -d

3. Start Frontend Service

npm install
npm run dev

Step 2: Configuring LLM Router in Eigent

Eigent allows you to configure custom API endpoints. We will set this up to point directly to your LLM Router instance.

1. Access Application Settings

Launch Eigent and navigate to the Home Page, then click on the Settings tab (usually a gear icon).

2. Locate Model Configuration

In the Settings menu, find and select the Models section. Scroll down to the Custom Model area and look for the configuration card (often labeled as Custom OpenAI Config or similar).

3. Enter Configuration Details

Click on the custom configuration card and fill in the following information to connect to LLM Router:

• API Key: Enter your LLM Router API Key (obtain from the LLM Router Dashboard). It usually starts with sk-router-.
• API Host / Base URL: Enter https://api.llmrouter.app/v1 (Make sure not to include /chat/completions at the end).
• Model Type: Enter the specific model you want Eigent to use, prefixed with the provider slug.
  • Example 1: anthropic/claude-3-5-sonnet (Best for complex desktop automation).
  • Example 2: openai/gpt-4o

Click Save to apply your changes.

4. Set as Default

Once saved, click the “Set as Default” button on the configuration card to ensure Eigent routes all its agentic tasks through LLM Router.

Step 3: The LLM Router Advantage

With the configuration complete, your Eigent agents are now powered by LLM Router.

Because Eigent executes tasks autonomously on your machine, the safety and cost-saving layers of LLM Router become critical:

• Zero-Trust PII Redaction: If an Eigent agent reads a local config file containing your AWS keys or Stripe tokens while trying to debug a project, LLM Router will automatically mask them into [TOKEN_REDACTED] before sending the context to Claude or GPT-4o.
• Context Pruning: Agents often “loop” and read the same files multiple times. LLM Router will intelligently compress the middle of these repetitive file reads, saving you thousands of tokens per task.
• Provider Resilience: If Anthropic goes down while Eigent is in the middle of a complex multi-step browser automation task, LLM Router will seamlessly fall back to OpenAI, preventing the agent from crashing and losing its progress.