🧠 Doubao GPT Engine Integration

Domestic LLM · Chinese Optimized · Text Generation

A base model access point tailor-made for localization and cost efficiency. Connects to ByteDance's Doubao AI engine via official endpoints, delivering silky-smooth context management and ultra-long text generation capabilities.

OpenClaw Team

🚀 Quick Install

Run the following command in your terminal to install:

npx clawhub install doubao-gpt

📊 Stats Overview

⭐ Stars	☁️ Total Calls	👥 Active Users	🎯 Stable Version
1.5k	9.42M	12,400	v3.5.1

🎛️ How It Works

Accessing overseas LLMs often faces network blockages and extremely high token costs. This module natively integrates with Volcengine, providing a perfect Chinese-optimized solution:

🇨🇳 Optimized Chinese Ecosystem: Unlike overseas models that rely on English-to-Chinese translation layers, Doubao's base model has extremely authentic localized expressions and contextual associations. Agents can use it to produce eloquent viral long-form articles or rigorous official documents.
⚡ Ultra-fast Streaming Output: Completely eliminates the stiff "waiting for hourglass" machine feel. The agent connects to the underlying SSE protocol, delivering generated natural language and even full code file snippets character by character, like real-time human typing.
💰 Minimal Cost: Using domestic optimized computing nodes, costs are so low compared to tens-of-dollars overseas bills that you can let AI run large-scale text extraction and summarization around the clock without worrying about budget.
📦 Session Context Window Management: Built-in efficient history conversation window trimming. When multi-turn Q&A exceeds the maximum context threshold, the engine automatically applies sliding window truncation to maintain the main reasoning thread.

🧭 Typical Use Cases

📝 Scenario 1: High-frequency Customer Service Cost Reduction

With hundreds of WeChat or private community groups needing AI-powered Q&A, overseas flagship model bills alone could bankrupt small businesses. doubao-gpt provides smart, responsive baseline Q&A at a fraction of the cost — with excellent performance in understanding internet slang and consumer intent.

📚 Scenario 2: Long-form Content Writing & Brainstorming

After capturing trending topics, the Agent can invoke the Doubao base model to produce 3,000+ word exploratory pieces in specific writing styles. The model's powerful continuous text generation returns beautifully structured, well-formatted Markdown content.

💻 Command Reference

After installation, you can let AI call these autonomously via conversation, or manually trigger operations from the CLI:

Basic Q&A mode — one-shot intent delivery and quick answers:

clawhub execute doubao-gpt \
  prompt="Summarize the core idea of 'The Three-Body Problem' Book 1 in 50 words"

Continuous dialogue with role configuration:

clawhub execute doubao-gpt \
  system_prompt="You are a senior Java architect with 20 years of experience" \
  prompt="Please explain JVM garbage collection's tri-color marking algorithm"

Flow control and diversity tuning:

clawhub execute doubao-gpt \
  prompt="Write a poem praising cyberpunk autumn" temperature=0.8 streaming=true

🛡️ Requirements & Authentication

🔑 Enterprise Endpoint Authorization: Ensure you've registered on the Volcengine console and obtained your API_KEY along with the specific Endpoint_ID for your Doubao model (typically in the format ep-xxxx...).
⚙️ Environment Variable Setup: Before using this module, pre-expose both parameters in your terminal (or .env): export VOLC_API_KEY="..." and export VOLC_ENDPOINT_ID="...".

🔗 View Official API Docs