qianwenai-wiki

Searches an offline snapshot of Qianwen AI Platform documentation to answer API integration questions.

733|99|Updated Dec 19, 2025
One-click install
npx skills add https://github.com/chujianyun/skills --skill qianwenai-wiki
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: qianwenai-wiki
Source: https://github.com/chujianyun/skills/tree/main/skills/knowledge/qianwenai-wiki
Command: npx skills add https://github.com/chujianyun/skills --skill qianwenai-wiki

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Developers integrating Qianwen AI Platform (DashScope) models often need accurate API parameters, SDK examples, and troubleshooting steps without browsing hundreds of online doc pages. This Skill packages 579 official documentation pages into an offline, full-text searchable knowledge base so answers come from verified sources instead of model memory.

Core Features & Use Cases

  • Offline Full-Text Search: A zero-dependency Python script searches 470 Markdown guides and 109 OpenAPI JSON schemas by keyword, model name, parameter, or error code, with directory filters for developer guides, API references, and billing docs.
  • Verified Snapshot with Integrity Checks: Every file records its source URL, size, and SHA-256 hash in SNAPSHOT.json, with secrets redacted during sync, so answers can cite exact official sources.
  • Controlled Refresh Workflow: A sync script re-downloads the official llms.txt index into a temporary directory, validates all pages, and only replaces the existing snapshot after explicit user confirmation.
  • Use Case: Ask how to authenticate the Qwen Realtime API or which parameters the Wanxiang text-to-video endpoint requires, and receive an answer grounded in the actual official documentation with cited file paths.

Quick Start

Ask the assistant to use qianwenai-wiki to show a Python example of calling a Qwen model through the OpenAI-compatible API with DASHSCOPE_API_KEY.

Frequently Asked Questions about qianwenai-wiki

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I search Qianwen AI Platform documentation offline?

Run the bundled search_docs.py script with 2-6 Chinese or English keywords, such as model names, API parameters, or error codes. Use the --prefix flag to restrict results to developer-guides, api-reference, resources, or token-plan directories.

What topics does the Qianwen AI Platform documentation cover?

The snapshot covers model selection, API keys, OpenAI-compatible and DashScope SDKs, text and multimodal generation, image and video generation, speech, Realtime API, embeddings, reranking, function calling, MCP, batch calls, billing, and Token Plan subscriptions.

Does the offline documentation include real-time pricing and model availability?

No. The snapshot reflects the fetch date recorded in SNAPSHOT.json and may lag behind the live site. Pricing, quotas, rate limits, and model availability should be verified against the official platform pages when internet access is available.

How do I update the offline documentation snapshot?

Run sync_docs.py, which downloads every page listed in the official llms.txt into a temporary directory and validates counts and hashes. Replacing an existing snapshot requires the explicit --replace flag after user confirmation.

What is the difference between OpenAI-compatible and DashScope native APIs?

They use different endpoints, request fields, and authentication patterns, and are not interchangeable. The documentation covers both, so the Skill identifies which interface you use and cites the matching reference pages rather than mixing fields.

Why does the documentation search return no results for my query?

No matches usually means the keywords do not appear verbatim in the docs. Retry with Chinese or English synonyms, older product names, specific model names, or a broader --prefix filter before concluding the capability is undocumented.