Skip to content

Releases

v0.1.3

Choose a tag to compare

@github-actions github-actions released this 23 Oct 15:52

This release adds DeepSeek web search and page fetching, and refines fetch fallback, output formatting, and Chinese help output.

Added

  • DeepSeek web search and page fetching, with deepseek search and deepseek fetch commands. Search returns summaries by default, with full page text available on request. deepseek fetch reads one URL and returns its title and Markdown body.

Fixed

  • fetch no longer switches services when the target page reports 401 or 451, matching the existing behavior for 404 and 410.
  • Model search answers report only cited sources.
  • Chinese help headings no longer carry a stray trailing space.

Changed

  • Standardize CLI labels, error details, and doctor diagnostics.
  • The default fetch.providers order now ends with deepseek. If this field was saved with config set, add deepseek to that list manually; the default order applies only when the field is unset.

本版本新增 DeepSeek 网页搜索与正文读取,并改进 fetch 的切换判定、输出格式与中文帮助输出。

新增

  • 新增 DeepSeek 网页搜索与正文读取,提供 deepseek search 与 deepseek fetch 命令。搜索默认返回摘要,也可获取完整正文;deepseek fetch 读取单个网页并返回标题与 Markdown 正文。

修复

  • 目标页面返回 401 或 451 时,fetch 不再尝试下一个服务,与既有的 404、410 行为一致。
  • 模型搜索回答只报告被引用的来源。
  • 中文帮助标题不再带有多余的空格。

变更

  • 统一 CLI 标签、错误详情与 doctor 诊断输出。
  • 默认 fetch.providers 序列末尾增加 deepseek。若曾用 config set 保存过该字段,需要手动把 deepseek 加入该列表;默认序列只在该字段未设置时生效。

v0.1.2

Choose a tag to compare

@github-actions github-actions released this 16 Oct 15:47

This release refines model search output and page reading behavior.

  • Remove inline <think> reasoning blocks from search answers.
  • Move model-written source lists into structured sources, and read the citations field returned by model APIs.
  • Switch to a backup model when a rate-limited request cannot be retried, without requiring the response to name the model as the cause.
  • Switch to another available model instead of waiting a retry delay that reaches half of the remaining time limit.
  • Stop fetch from trying the next service when the target page reports 404 or 410; other page failures such as 403 still switch.
  • Order fetch page fields consistently in JSON output.

Upgrade with npm:

npm install -g unisearch-cli@0.1.2

本版本改进模型搜索的输出,以及网页读取的行为。

  • 移除 search 回答中的内联 <think> 推理块。
  • 将模型自写的来源列表移入结构化的 sources,并读取模型 API 返回的 citations 字段。
  • 请求被限流且无法重试时改用备用模型,不再要求响应明确标注为模型自身的原因。
  • 存在其他可用模型时,若重试等待达到剩余时限的一半,改用该模型而非继续等待。
  • 目标页面返回 404 或 410 时,fetch 不再尝试下一个服务;403 等其他页面失败仍会切换。
  • fetch 的 JSON 输出改用一致的页面字段顺序。

使用 npm 升级:

npm install -g unisearch-cli@0.1.2

v0.1.1

Choose a tag to compare

@github-actions github-actions released this 11 Oct 06:32

This release improves diagnostics, Markdown output, Skill synchronization, and Jina search behavior.

  • Improve doctor output, including Tavily quota handling for free accounts and clearer verbose diagnostics.
  • Improve Skill status and synchronization for directory symlinks, with more compact target output.
  • Separate Markdown execution details from retrieved content and simplify generated result formatting.
  • Render link-only search results as compact ordered lists instead of empty heading sections.
  • Make Jina search metadata-only by default; use --include-text when page text is needed, or use jina fetch for selected URLs.

Upgrade with npm:

npm install -g unisearch-cli@0.1.1

本版本改进诊断、Markdown 输出、Skill 同步和 Jina 搜索行为。

  • 改进 doctor 输出,包括免费 Tavily 账户的额度判断和更清晰的详细诊断信息。
  • 改进目录软链接下的 Skill 状态检查与同步,并压缩目标输出排版。
  • 将 Markdown 执行详情与检索正文分开,简化自动生成的结果格式。
  • 只有链接的搜索结果改用紧凑的有序列表,避免空的标题区块。
  • Jina 搜索默认只返回元数据;需要正文时使用 --include-text,或对选定 URL 使用 jina fetch。

使用 npm 升级:

npm install -g unisearch-cli@0.1.1

v0.1.0

Choose a tag to compare

@github-actions github-actions released this 10 Oct 13:38

Initial public release of UniSearch, a unified retrieval CLI for AI agents.

  • Search through model APIs with search, with support for extra search sources.
  • Read page content with fetch and discover site links with map.
  • Look up library and API documentation with Context7, and explore documentation in public GitHub repositories with Zhipu MCP.
  • Configure providers and install the bundled Skill for supported agents with setup; inspect local settings with status.
  • Use Chinese or English CLI messages and JSON or Markdown output.

Requires Node.js 24 or later and npm on Linux, macOS, or Windows.

Install v0.1.0, then run setup in an interactive terminal:

npm install -g unisearch-cli@0.1.0
unisearch --version
unisearch setup

v0.1.0 documentation: Getting started · Configuration


UniSearch 首次公开发布,提供面向 AI Agent 的统一检索 CLI。

  • 使用 search 通过模型 API 检索信息,并支持额外搜索来源。
  • 使用 fetch 读取网页正文,使用 map 发现站点内的链接。
  • 通过 Context7 查询库与 API 文档,通过智谱 MCP 查阅公开 GitHub 仓库的文档。
  • 使用 setup 交互配置服务、为支持的 Agent 安装随包 Skill,使用 status 检查本地配置。
  • 提供中英文 CLI 界面,以及 JSON 和 Markdown 输出。

支持 Linux、macOS 和 Windows,需要 Node.js 24 或更高版本及 npm。

安装 v0.1.0,并在交互终端中完成配置:

npm install -g unisearch-cli@0.1.0
unisearch --version
unisearch setup

v0.1.0 文档:开始使用 · 配置文档