Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Volcengine Launches Doubao Search Service, Opens to Agent API Calls

Volcengine, ByteDance cloud arm, has officially launched the Doubao Search service, extracting the search capability behind the Doubao app and offering it to enterprises and developers via API, MCP, and Skill interfaces. The service returns search results enriched with authority ratings, publication timestamps, text summaries, and original Markdown excerpts, allowing AI agents to assess information credibility and extract key content without reading full web pages.

Published
火山引擎正式上线豆包搜索服务,开放Agent调用
Image source: champsorchumps.us

Volcengine, ByteDance cloud computing arm, has officially launched the Doubao Search service, extracting the search capability that powers the Doubao consumer app and making it available to enterprises and developers via API, MCP, and Skill integrations. The move marks another significant step in ByteDance AI infrastructure strategy on its cloud platform.

Unlike conventional search APIs, Doubao Search returns not only web links but also structured metadata for each result: authority grading, publication timestamp, a targeted summary aligned with the query, and original Markdown excerpts. This means an agent receives not just a list of URLs but the key fields needed to decide whether and how to use each piece of information.

The design philosophy reflects a fundamental shift in search for the agent era. Large language models excel at comprehension, summarization, and reasoning, but face inherent knowledge freshness boundaries. When agents encounter breaking events, new product versions, or changing prices, they need search for up-to-date information. If search only returns URLs, the agent must still open, read, and extract from each page — Doubao Search aims to eliminate this intermediate processing overhead.

In hands-on testing documented by QbitAI, using OpenClaw with the Doubao-Seed-Evolving model, Doubao Search demonstrated strong performance across multiple scenarios. In a real-time query about the Fields Medal, it returned results from government sources, official science media, and the Simons Foundation, each tagged with source type and authority level. When Kimi K3 was open-sourced in the early hours, Doubao Search rapidly surfaced weight files, license details, repository links, and deployment information, allowing the agent to process these directly without re-reading full articles.

On systematic benchmarks, Doubao Search boosted SimpleQA (factual accuracy) scores by 70% compared to the no-search baseline. It also achieved top gains on FreshQA (recency-focused questions), BrowseComp-ZH (Chinese browsing and deep retrieval), and Xbench-2505 (Chinese deep search and professional scenarios).

Doubao Search is now officially available to enterprises and developers. Individual users receive 500 free searches per month, with pay-as-you-go and subscription plans available beyond that. Developers can install the byted-web-search Skill via a single command in OpenClaw or other Claw environments, then create an API key from the Volcengine console.

Vertical coverage includes financial research, business intelligence, product insights, content marketing, and office automation. Results support filtering by industry, time range, domain, and authority level, giving the upstream agent flexible control over how recent, how broad, and which sources to prioritize.

As agents grow more capable, search is evolving from a tool for information retrieval into an essential gateway connecting agents to the external world. Doubao Search represents ByteDance pivot from C-end search capabilities to B-end agent infrastructure, a move that could reshape the AI agent development ecosystem in China.

Why it matters

Doubao Search transforms ByteDance consumer search capability into enterprise-grade agent infrastructure, potentially becoming a foundational layer for AI agent development in China.

VolcengineDoubaoSearchAgentByteDance
Back to realtime news

Nearby Updates

All

07/29, 17:50

Caixin: China Backs AI Token Business Innovation at Policy Level

China is offering policy-level support for AI token business innovation, according to a Caixin Global business brief on July 29. The development signals a more favorable regulatory posture toward token-economy business models within the AI industry.

07/29, 17:01

Alibaba's Qoder Launches Real-Time Voice Agent Qoder Voice for AI-Powered Coding

Alibaba's AI coding assistant Qoder has launched Qoder Voice, China's first real-time voice-interactive agent for programming. Powered by the Qwen-Audio-3.-Realtime full-duplex model, it allows developers to create tasks, discuss solutions, and execute code through natural speech.

07/29, 16:52

Google Launches Gemini Spark in India: An AI Agent That Works While You Sleep

Google has officially launched Gemini Spark in India, an AI agent product capable of autonomously executing tasks while users sleep. According to News18, users can assign tasks before leaving work and receive completed results by the next morning, representing Google's latest push in the AI agent space.

07/29, 16:50

AMD Unveils MI-Series GPUs, EPYC CPUs, and Helios Rack Systems to Challenge Nvidia

AMD has launched its MI-Series GPUs, EPYC CPUs, and Helios rack systems, mounting a direct challenge to Nvidia's dominance in AI chips. According to EE Times Asia, the Helios rack system integrates AMD's GPUs and CPUs into a unified solution, marking AMD's transition from chip vendor to full-stack AI infrastructure provider.