Outlit
概览
作者提供的说明(英文原文)
Ask about any customer and get one matched record across CRM, billing, support and product data.
PopNx MCP 的三大支柱
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
- 用户评价i暂无评价评价功能将在后续阶段与账户功能一起开放
- 我们的基础检查i3 项中通过 3 项依据公开信息检查;尚未扫描代码
- 使用情况i非常新★ –
它能帮你做什么
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
- 连接方式: 在云端运行
- 传输方式: Streamable HTTP
- 分类: 营销与社交媒体
安装
选择你使用的 AI 应用,然后复制设置。
安装之前
- 基础检查:3 项中通过 3 项
- 在作者的服务器上运行:你向它提出的内容会经过他们的服务器。
- 注册中心没有列出需要填写的密钥。
你需要
- 在云端运行无需安装任何东西,通过互联网连接
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
适用于网页版、桌面版和手机版 Claude:
- 打开 Settings → Connectors
- 点击 Add custom connector
- 粘贴下面的地址,然后点击 Add
- 如果要求登录,请登录作者的服务
https://mcp.outlit.ai/mcp在终端中运行这条命令
claude mcp add --transport http outlit https://mcp.outlit.ai/mcp或者自己复制设置
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"outlit": {
"url": "https://mcp.outlit.ai/mcp"
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
或者自己复制设置
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"outlit": {
"type": "http",
"url": "https://mcp.outlit.ai/mcp"
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
ChatGPT 只能连接互联网上的服务器:
- 打开网页版 ChatGPT → Settings → Connectors (Apps)
- 点击 Create 或 Add(部分套餐需要先在 Advanced 中开启 Developer mode)
- 粘贴下面的地址,起个名字,然后保存
- 如果要求登录,请登录作者的服务
https://mcp.outlit.ai/mcp菜单名称可能因套餐和版本不同而略有差异。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它能做什么
服务器要求先登录,才会列出它的工具。
基础检查i
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月8日
- ✓ 通过
作者与仓库一致
注册名称已通过域名 outlit.ai 验证(注册中心会核实域名所有权)
- … 尚未检查
仍在维护
我们还没有从 GitHub 读取这个仓库
- … 尚未检查
有许可证
我们还没有从 GitHub 读取这个仓库
- ✓ 通过
端点有响应
在 1403 毫秒内响应(HTTP 401)
- ✓ 通过
使用 OAuth 登录
声明支持 OAuth 登录,你的 AI 应用可以带你去登录
- … 尚未检查
工具描述中没有隐藏指令
尚未检查;我们会在检查端点时读取工具列表
与“营销与社交媒体”中的其他服务器比较i
- 星标数多于此分类中其他 3602 个服务器的 0%
- 发布到注册中心的时间比此分类中其他 3602 个服务器的 89% 更新
- 基础检查结果等于或优于此分类中其他 3602 个服务器的 100%
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
常见问题
Outlit 安全吗?
它在 3 项适用的基础检查中通过了 3 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
它需要哪些密钥或代码?
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
它在我的电脑上运行还是在云端?
它在作者的云端运行。通过互联网连接,无需安装,但你发送的内容会经过作者的服务器。
它还在维护吗?
GitHub 仓库最后更新于–,最新版本(1.1.0)于3天前进入注册中心。
哪些应用可以使用它?
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,所以接受远程 MCP 服务器的应用也可以使用,例如开启开发者模式的 ChatGPT。
有什么可以替代它?
同一分类中有:MCP YouTube Transcript, Google Analytics MCP, BibiGPT YouTube Video Summarizer
替代选择
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:↓ 每周下载量i★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Google Analytics MCP
GA4 MCP server for AI agents: schema discovery, server-side aggregation, analysis-ready defaults.
BibiGPT YouTube Video Summarizer
AI video summarizer: summarize and transcribe YouTube, Bilibili, TikTok and podcast videos.
作者的 README
打开 README
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
Outlit SDK
Outlit helps customer-success teams protect and grow long-tail revenue. It connects customer data, monitors for churn risk and expansion opportunities, and brings attention items with supporting evidence to your team.
Start with the product guides for guided setup and your first attention item. For programmatic access, use the developer and agent guide.
This repository contains the public SDK and developer integration packages for Outlit: browser, server, CLI, tool contract, and Pi packages for sending customer signals to Outlit and querying customer context from agent workflows. It is not the hosted remote MCP server implementation; use the canonical Outlit-owned discovery endpoints below for MCP, API, and agent metadata.
Packages
This monorepo contains these packages under the @outlit scope:
- [@outlit/core](./packages/core) - Core SDK functionality and base client
- [@outlit/browser](./packages/browser) - Browser-specific SDK with automatic page view tracking
- [@outlit/node](./packages/node) - Node.js SDK for server-side event tracking
- [@outlit/cli](./packages/cli) - CLI for Outlit customer intelligence
- [@outlit/tools](./packages/tools) - Customer intelligence tool contracts and client helpers for API and agent integrations
- [@outlit/pi](./packages/pi) - Pi package with Outlit customer intelligence tools and skill guidance
Agent and Crawler Discovery
Use these canonical resources for citations, schema-driven clients, and agent setup instead of copying generated specs or server metadata into this repository:
| Surface | Canonical URL | Purpose |
|---|---|---|
| Documentation | https://docs.outlit.ai | Product guides plus SDK, CLI, API, MCP, and customer context documentation |
| Docs index for agents | https://docs.outlit.ai/llms.txt | Machine-readable map of documentation pages |
| Product resource index | https://www.outlit.ai/llms.txt | Agent-facing map of SDK packages, docs, API contracts, MCP, CLI, Pi, and skills |
| OpenAPI spec | https://docs.outlit.ai/openapi.json | Canonical OpenAPI contract for public API and ingest surfaces |
| API catalog | https://www.outlit.ai/.well-known/api-catalog | Linkset for API, MCP, OAuth, docs, and support discovery |
| AI catalog | https://www.outlit.ai/.well-known/ai-catalog.json | Agentic Resource Discovery catalog for Outlit API, MCP, skills, and SDK resources |
| MCP Registry listing | https://registry.modelcontextprotocol.io/v0.1/servers?search=ai.outlit/outlit | Official MCP Registry search surface for ai.outlit/outlit |
| MCP server metadata | https://mcp.outlit.ai/.well-known/mcp/server.json | Runtime metadata for the hosted remote MCP server |
| MCP server card | https://mcp.outlit.ai/.well-known/mcp/server-card.json | Runtime discovery card for the hosted remote MCP server |
| MCP docs | https://docs.outlit.ai/ai-integrations/mcp | Connect remote MCP clients with workspace URLs and OAuth |
| Agent skills | https://docs.outlit.ai/ai-integrations/skills | Official outlit and outlit-sdk skill installation guidance |
The hosted MCP server and OAuth metadata live on mcp.outlit.ai; this SDK repo is the public package and developer integration surface.
Installation
Choose the package that matches the integration surface:
| Package | Install | Use when |
|---|---|---|
@outlit/browser | npm install @outlit/browser | Browser apps, React, Next.js, Vue, Nuxt, SvelteKit, Angular, Astro, and script-tag tracking |
@outlit/node | npm install @outlit/node | Node.js servers, API routes, jobs, webhooks, CLIs, desktop main processes, and native JavaScript runtimes |
@outlit/core | npm install @outlit/core | Lower-level custom SDK implementations that do not need browser or Node runtime helpers |
@outlit/cli | npm install -g @outlit/cli | Terminal access to Outlit customer intelligence and setup workflows |
@outlit/tools | npm install @outlit/tools | Custom API or agent integrations that need typed Outlit tool gateway contracts and client helpers |
@outlit/pi | npm install @outlit/pi | Pi agents that need Outlit customer intelligence tools and skill guidance |
| Rust crate | cargo add outlit | Rust backends, CLIs, and Tauri backends |
Tracking SDK examples:
# For browser applications
npm install @outlit/browser
For Node.js applications
npm install @outlit/node
For custom implementations
npm install @outlit/core
Quick Start
Browser
import { Outlit } from '@outlit/browser'
const outlit = new Outlit({ publicKey: 'pk_xxx', trackPageviews: true, trackForms: true, })
// Identify a user outlit.user.identify({ email: 'user@example.com', traits: { name: 'John Doe' }, customerId: 'cust_123', // Your app's account/workspace/customer ID customerTraits: { plan: 'pro' }, })
// Track events outlit.track('button_clicked', { button_id: 'signup', page: '/homepage', })
// Track meaningful product activity. Core derives lifecycle stages from // ordinary events, including your selected activation event. outlit.track('onboarding_completed', { flow: 'self_serve' })
Using the singleton API
import { init, track, user } from '@outlit/browser'
// Initialize once at app startup init({ publicKey: 'pk_xxx' })
// Then use anywhere track('page_viewed', { page: '/home' }) user().identify({ email: 'user@example.com', customerId: 'cust_123', // Your app's account/workspace/customer ID }) track('subscription_upgraded', { plan: 'pro' })
Using with React
import { OutlitProvider, useOutlit } from '@outlit/browser/react'
// Wrap your app function App() { return (
) }
// Use in components function MyComponent() { const { track } = useOutlit()
return ( track('onboarding_completed')}> Click me
) }
Node.js
import { Outlit } from '@outlit/node'
const outlit = new Outlit({ publicKey: 'pk_xxx', })
// Track server-side events (requires identity) outlit.track({ customerId: 'cust_123', // Your app's account/workspace/customer ID eventName: 'api_request', properties: { endpoint: '/api/users', method: 'GET', status: 200, }, }) // customerId-only track events are valid immediately. // When you later call identify() with the same customerId and an email, // Outlit can link that account/workspace to the customer resolved from email.
// Identify a user outlit.user.identify({ email: 'user@example.com', traits: { plan: 'pro' }, customerId: 'cust_123', // Your app's account/workspace/customer ID customerTraits: { plan: 'pro' }, })
// Track ordinary product events. Billing status comes from verified // integrations such as Stripe, not authoritative SDK commands. outlit.track({ customerId: 'cust_123', eventName: 'subscription_upgraded', properties: { plan: 'pro' }, })
// Flush events before shutdown await outlit.flush()
Features
- Modern TypeScript - Full TypeScript support with type definitions
- Tree-shakeable - Optimized bundle size with dual ESM/CJS exports
- Event Queueing - Automatic batching and flushing of events
- Multi-platform - Separate packages for browser and Node.js
- Auto-tracking - Automatic page view tracking in browser
- Middleware Support - Easy integration with Express and similar frameworks
- Persistent Identity - User and anonymous ID persistence
- High Performance - Minimal overhead and efficient batching
- Type Safe - Full TypeScript support with strict types
Examples
- [Pi agents](./examples/pi-agents) - Build customer intelligence agents in Pi with
@outlit/pi
Development
This project uses a modern monorepo setup with the following tools:
- [Bun](https://bun.sh/) - Fast all-in-one JavaScript runtime and package manager
- [Turbo](https://turbo.build/) - Build system for monorepo orchestration
- [TypeScript](https://www.typescriptlang.org/) - Type-safe JavaScript
- [tsup](https://tsup.egoist.dev/) - Fast TypeScript bundler
- [Biome](https://biomejs.dev/) - Fast linter and formatter
- [Playwright](https://playwright.dev/) - End-to-end testing
- [Changesets](https://github.com/changesets/changesets) - Version management and changelogs
Setup
# Install dependencies
bun install
Build all packages
bun run build
Run tests
bun run test
Lint code
bun run lint
Type check
bun run typecheck
Format code
bun run format
Project Structure
outlit-sdk/
├── .github/workflows/ # CI/CD workflows ├── examples/ │ └── pi-agents/ # Example Pi agents using @outlit/pi ├── packages/ │ ├── browser/ # Browser SDK with React bindings │ ├── cli/ # Outlit CLI │ ├── core/ # Shared types and utilities │ ├── node/ # Node.js SDK │ ├── pi/ # Pi package for Outlit tools │ ├── tools/ # Customer intelligence tool contracts │ └── typescript-config/ # Shared TypeScript configs ├── package.json # Root package with workspace config ├── bun.lock # Bun lockfile ├── turbo.json # Turbo build configuration └── biome.json # Biome linter/formatter config
Creating a Changeset
When making changes that should be released, create a changeset:
bunx changeset
This will prompt you to:
- Select which packages are affected
- Choose the version bump type (patch, minor, major)
- Write a description of the change
The changeset file will be committed with your PR and used to generate changelogs on release.
CI/CD
Workflows
- CI (
ci.yml) - Runs on PRs: lint, typecheck, build, test - Release (
release.yml) - Runs on main: publish canary to npm + CDN, create version PR or publish stable releases - Verify CDN Auth (
verify-cdn-auth.yml) - Manual only: checks Workload Identity Federation exchange and CDN read access; the optionalverify_writeinput writes then deletes one tiny test object under__compliance_checks/
Stable releases are intentionally separate from SDK source merges. See docs/release-coordination.md for the Version Packages checklist and the required Core-production-before-stable-SDK order for coordinated contracts.
Required Secrets
For maintainers setting up the repository:
| Secret | Description |
|---|---|
NPM_TOKEN | npm access token with publish permission for @outlit scope |
CDN deploys authenticate to Google Cloud via Workload Identity Federation instead of a stored service account key. Maintainers configure these repository variables:
| Variable | Description |
|---|---|
GCP_WORKLOAD_IDENTITY_PROVIDER | Full WIF provider resource name (projects/ /locations/global/workloadIdentityPools/ /providers/ ) |
GCP_SERVICE_ACCOUNT | Deploy service account email with Storage Object Admin on cdn.outlit.ai bucket |
Creating NPM_TOKEN
- Go to npmjs.com and sign in
- Navigate to Access Tokens → Generate New Token → Granular Access Token
- Set permissions: Read and write for
@outlitpackages - Copy the token and add as
NPM_TOKENsecret in GitHub
Creating the GCP Workload Identity setup
- Go to Google Cloud Console
- Navigate to IAM & Admin → Service Accounts
- Use the deploy service account (
github-actions-cdn-deploy) - Grant "Storage Object Admin" role on the
cdn.outlit.aibucket - Create a dedicated Workload Identity Pool for this repo with an OIDC provider for GitHub (issuer `https://token.actions.githubusercontent.
只显示了开头部分;其余内容请在仓库中阅读