npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

chaite

v1.15.2

Published

core for chatgpt-plugin and karin-plugin-chatgpt

Readme

node-chaite

用于chatgpt-plugin v3和karin-plugin-chaite的公用依赖。

只包含核心功能,尽可能轻量,不包含数据库和其他存储。

Claude thinking 兼容

Claude 4.6 及更新的自适应模型会自动使用 thinking.type=adaptive,通过已有的 reasoningEffort 设置 output_config.effort;Haiku 4.5、Sonnet 4.5 等旧模型仍使用 reasoningBudgetTokens。自适应请求不再发送手动预算或温度参数。手动预算必须至少为 1024,且小于 maxToken。

若网关重命名了模型,无法从名称判断版本,可在 SendMessageOption 中显式覆盖:

{
  model: 'my-haiku-alias',
  enableReasoning: true,
  claudeThinkingMode: 'adaptive',
  claudeEffort: 'high', // low / medium / high / xhigh / max
  maxToken: 8192
}

Claude 返回的 thinking 签名、redacted thinking 和工具调用块会随历史原样保留。 不要删除 provider_context 内容或修改其中的签名;它们用于后续工具轮次的协议回传, 不用于显示。新模型不支持 assistant prefill,请以用户消息或工具结果结束请求。

OpenAI / Gemini 协议兼容

  • OpenAI 推理模型会省略不支持的温度参数;可用 openaiReasoningEffort 指定 none/minimal/low/medium/high/xhigh/max,具体可用值取决于模型。
  • OpenAI 官方请求不发送 thinking_budget_tokens。llama-server 等明确支持该扩展的 兼容服务,改用 openaiThinkingBudgetTokens 显式设置,不再复用所有预设里的通用预算。
  • 未识别为 OpenAI 官方模型的兼容服务默认使用 max_tokens。可用 openaiMaxTokenField: 'max_tokens' | 'max_completion_tokens' 覆盖网关需求。
  • OpenAI 当前适配器使用 Chat Completions。GPT-6 Astra、GPT-6.1 Sol 的工具调用需要 Responses API,配置这些模型与工具时会给出明确错误,不能通过换 SDK 参数实现。
  • Gemini 3/4 使用 thinkingLevel,2.5 使用 thinkingBudget。可用 geminiThinkingLevel: 'minimal' | 'low' | 'medium' | 'high' 覆盖自动映射;不支持关闭 thinking 的模型使用其最低支持等级,2.5 Pro 使用最低预算 128。
  • Gemini 原始 parts、空签名、函数调用顺序和服务端工具上下文会完整保留。 两个适配器均规范化图片 MIME/Base64,toolChoice=none 会禁用工具。

协议回归测试无需真实 API Key:先运行 pnpm build,再运行 pnpm test:adapters, 使用本机临时 HTTP 服务验证构建产物与真实 SDK 的两轮工具调用(含 OpenAI 流式并行调用)。

预计功能

  • [x] 主要LLM适配器
    • [x] OpenAI
      • [x] 对话
      • [x] 视觉/音频
      • [x] 工具
      • [x] 嵌入
    • [x] Gemini
      • [x] 对话
      • [x] 视觉
      • [x] 工具
      • [x] 嵌入
    • [x] Claude
      • [x] 对话
      • [x] 视觉
      • [x] 工具
    • [ ] 消息转换器 custom hook
  • [x] 处理器
    • [x] 前处理器
    • [x] 后处理器
  • [ ] 其他模型接口抽象
    • [ ] 生图
    • [ ] 音频
  • [x] 触发器
    • [x] Cron触发器
    • [] Event
  • [x] 抽象历史管理
  • [x] 知识库
    • [x] 抽象向量库构建
    • [x] 抽象检索
  • [x] 辅助工具
    • [x] 多模态文件解析
      • [x] 管理器 documentParserManager
      • [x] txt
      • [x] word
      • [x] pdf
  • [x] 预设/工具管理
    • [x] 本地热插拔
    • [x] 云端共享
    • [x] 点对点共享
    • [x] 云端实现
  • [x] 渠道管理