Access GPT, Claude, and Gemini in LobeChat with DGrid AI Gateway

LobeChat is an open-source, production-grade AI Agent Workspace (starred by 60k+ developers on GitHub) designed for seamless integration of large language models (LLMs), multi-modal interactions, and extensible workflows. Beyond core chat functionality, it offers a comprehensive ecosystem including Knowledge Base with RAG capabilities, one-click MCP (Model Context Protocol) plugin installation, a curated Agent Marketplace, desktop/mobile cross-device support, and self-hosting flexibility.

This guide systematically outlines the workflow for acquiring a DGrid API key and implementing DGrid AI Gateway service within LobeChat, with a focus on security best practices and technical validation.

Prerequisites

1. A Web3 wallet (e.g., MetaMask) for DGrid account authentication and API key generation.
2. An active LobeChat instance (web-based via lobechat.com , self-hosted, or desktop app — see GitHub Deployment Guide for setup).
3. A secure credential management tool to store sensitive DGrid API credentials.
4. Network access to DGrid’s infrastructure: API Key Console and RPC endpoint (https://api.dgrid.ai/v1).

What Is DGrid AI Gateway?

DGrid AI Gateway is a unified inference gateway for decentralized AI.

Instead of integrating with each provider separately, applications send requests to DGrid. The network then routes those requests to the appropriate model provider.

The architecture looks like this:

AI Client (LobeChat)
        │
        ▼
   DGrid AI Gateway
        │
 ┌──────┼─────────────┐
 ▼      ▼             ▼
OpenAI  Anthropic     Google
 GPT      Claude      Gemini

This design allows applications to switch between models without changing infrastructure.

Key Benefits

• Access to 200+ AI models
• OpenAI-compatible API format
• Decentralized model marketplace
• Lower inference costs
• Flexible model switching

Step 1: Get a DGrid API Key

The API key generation process aligns with the procedure detailed in the AnythingLLM integration guide. For technical consistency, follow these critical steps:

1. Navigate to the DGrid API Key Console (https://dgrid.ai/api-keys ).
2. Authenticate via your Web3 wallet.
3. Generate a new API key:

• Click Create New Key to initiate the generation process.
• Assign a descriptive, context-rich label (e.g., “LobeChat-RPC”) to facilitate access control and audit logging.
• Optional but recommended: Configure a credit limit or expiration timestamp to mitigate financial and security risks associated with unauthorized usage.
• Confirm key creation by selecting ​Create​.

4. Secure the API key immediately: The credential is displayed only once post-generation. Copy it to your secure credential manager — do not store it in plaintext, version control systems (e.g., Git), or shared environments.

Critical Security Advisory

Treat DGrid API keys as sensitive authentication tokens. Unauthorized access may result in unauthorized charges, data breaches, or service misuse. Implement the following safeguards:

• Restrict key access to authorized personnel only.
• Avoid transmitting keys via unencrypted channels (e.g., email, instant messaging).
• Regularly rotate keys (recommended every 90 days) via the DGrid API Console.

Configuration of DGrid AI Gateway in LobeChat

Leverage DGrid’s OpenAI protocol compatibility to configure the service in LobeChat. Adhere to the step-by-step workflow below for optimal reliability:

Access LobeChat and Authenticate:

• Launch a web browser and navigate to the LobeChat official website: https://lobechat.com/ .
• Select Get Started to initiate the authentication process. Use a supported identity provider to log in to your LobeChat account.

Navigate to AI Service Provider Settings:

• Post-authentication, locate and click your user avatar in the top-left corner of the LobeChat interface to expand the user menu.
• From the dropdown menu, select Settings to access the configuration dashboard.
• Within the Settings panel, identify and select AI Service Provider — this module governs all LLM integration parameters.

Enable and Configure OpenAI-Compatible Integration

• On the AI Service Provider page, scroll to the Disabled Providers section and locate the Open AI option.
• Click Open AI to expand its configuration pane (inactive state by default).
• Input the required DGrid credentials and endpoints:
  • ​API Key​: Paste the DGrid API key stored in your credential manager.
  • ​API Proxy URL​: Specify DGrid’s official RPC endpoint: https://api.dgrid.ai/v1 .

Activate the integration in the top-right corner of the configuration pane.

Important: OpenAI-Compatible Does NOT Mean OpenAI Only

DGrid AI Gateway uses an ​OpenAI-compatible API format​. This means most AI clients can connect immediately without custom integrations. However, ​OpenAI-compatible does not mean OpenAI-only​. The API format simply defines the ​request structure​, not the model provider. For example, the following models can all be used through the same endpoint:

All requests go through the same API endpoint:

https://api.dgrid.ai/v1

The only thing that changes is the ​model name​.

Troubleshooting

Claude Models Not Appearing in LobeChat

Some AI clients only list predefined OpenAI models.

If Claude models do not appear:

1. Look for Custom Model
2. Enter the model name manually
3. Search and enter model information (including context window and standard output limit).

Example:

Model ID: claude-3-sonnet
Model Display Name: claude 3 sonnet
Maximum Context Window: 200k

Save the configuration and refresh.

Connection Errors

Verify the following:

• API key is valid
• Base URL is correct

https://api.dgrid.ai/v1

• Your network allows outbound requests

Why Use DGrid Instead of Direct Model APIs?

Traditional integrations require connecting to multiple providers individually.

Example:

DGrid simplifies this into a single endpoint:

DGrid AI Gateway → Multiple AI Providers

Benefits include:

• simpler integration
• multi-model flexibility
• cost optimization
• decentralized inference infrastructure

Conclusion

Integrating DGrid’s smart RPC service with LobeChat combines the best of both platforms: DGrid’s open, low-cost, community-powered AI network and LobeChat’s feature-complete AI Agent Workspace. This integration empowers users to leverage decentralized LLMs across LobeChat’s most powerful workflows — from private document analysis (RAG) to multi-modal interactions and plugin-enhanced tasks — without sacrificing flexibility, privacy, or ease of use.

LobeChat’s active development and vibrant community ensure ongoing compatibility and feature expansion, making it an ideal choice for developers, researchers, and teams seeking a production-ready AI toolchain.

For advanced configuration (e.g., custom model routing, rate limiting) or technical support:

• DGrid Official Documentation: https://docs.dgrid.ai/
• LobeChat Technical Documentation: https://lobehub.com/docs/usage/start
• LobeChat Self-Hosting Guide: https://lobehub.com/docs/self-hosting/start