China Public Tender & Bid Intelligence API

The Problem to Solve

Public tender and bid data is a vital fuel for market intelligence, sales lead generation, and competitive analysis. However, this data is scattered across numerous national and provincial platforms with inconsistent formats, making manual cleaning and correlation extremely time-consuming. Engineers and data analysts often struggle with: How to efficiently obtain the full lifecycle of bids for a specific region, industry, or company? How to transform unstructured announcement text into structured, analyzable fields? How to quickly identify potential opportunities from planned projects? Manual processing is not only inefficient but also risks drawing skewed conclusions by overlooking data boundaries and pagination rules.

Core Capabilities & Workflow

This skill provides an API interface suite that encapsulates access to Shibo Tech's tender and bid database. The core workflow can be summarized as:
1. Authentication & Configuration: Configure access via the BBIAO_API_KEY environment variable. Supports both direct HTTP calls and the MCP protocol.
2. Multi-Dimensional Search:
* Use searchProjectApi for structured search, combining queries with keyword (space for AND, | for OR), inClueKW (must include), and excludeKW (exclude).
* Filter projects by company roles such as partAName (Client), partBName (Winner).
* Call industryReasoning to map natural language industry terms to standard industry codes.
3. Deep Retrieval & Analysis:
* Based on the id from search results, call getZTBStructreDetail to get structured details like budget, winning bid amount, and contact information.
* Use getProjectByProjectNumber to link all announcements under the same project number, forming a project timeline.
* Analyze a specific company's bidding activity and upstream/downstream networks using the corporate profile interfaces (e.g., companyProfileSummary).
4. AI Augmentation:
* Via aiSearchSubmitPolling, automatically rewrite natural language queries (e.g., "hospital medical bed procurement in East China over the last 6 months") into structured search conditions.
* Employ SearchProjectForAI for more flexible, natural-language-friendly searches.
* Use ztbAiStructureInfo to extract structured fields like subjects, amounts, and timestamps from announcement text.

Boundaries & Key Considerations

• Data is Not Exhaustive: All search interfaces support pagination (e.g., searchProjectApi has a limit of 50 results per page). Do not calculate market share or conclude "no relevant projects exist" without retrieving all pages.
• Role Identification: The companyName field only indicates a company's involvement and does not directly confirm it as the client or winner; its role must be verified via the structured details.
• Data Timeliness: The parameter for publication time differs between the planned project search (publishtime) and its attachment list (publishTime); attention is required when making calls.
• Compliance in Use: Contact information retrieved via companyProfileContacts must be used in compliance with relevant privacy and communication regulations.