Caelario

io.github.chunkydonut21由 GitHub 账号 chunkydonut21 发布v0.1.8更新于 Nov 8, 2026

MCP server for Instagram, TikTok, YouTube, X and Facebook data extraction and automation.

概览

AI 生成的概览

Caelario 是一个本地 MCP 服务器,为助手提供类型化工具,用于提取和自动化 Instagram、TikTok、YouTube、X 与 Facebook 的社交媒体数据。

可以做什么
对五个平台执行即时和持久化提取,获取个人资料、帖子等社交媒体数据,并支持结果分页和每次任务的额度估算。搜索和发现内容与账号,并管理批量任务和定时计划以实现周期性采集。通过 Caelario 工作区访问已保存的结果和用量工作流。在运行任务前发现可用能力。
适用场景
当你希望助手在无需自行处理各平台凭据的情况下,从 Instagram、TikTok、YouTube、X 或 Facebook 拉取公开社交媒体数据,或设置周期性提取任务时使用。
运行要求
通过 npm 包 caelario-mcp 以 stdio 在本地运行,需要 Node.js 20 或更高版本。需要 Caelario 工作区 API 密钥,通过环境变量 CAELARIO_API_KEY 提供;安装程序会提示输入并在用户范围注册服务器。还需要 Caelario 账号以及对托管 API 的网络访问。
安装前请注意
服务器使用可撤销的 Caelario 工作区 API 密钥(CAELARIO_API_KEY),应妥善保密。提取、批量和定时任务会消耗 Caelario 额度,可能产生计费,请查看服务商的计费说明。任务会向 Caelario 的专有托管 API 发送请求,平台身份由 Caelario 管理而非由你管理。定时计划和批量任务会创建周期性工作,可能持续消耗额度。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Caelario,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Desktop only,通过 STDIO。 STDIO 服务会启动本地进程,因此需要 SourceWeft 桌面宿主。

其他 MCP 客户端

参照 仓库 中的启动说明。

README

Caelario MCP

MCP server for social media data from Instagram, TikTok, YouTube, X and Facebook. Run extractions, search and discovery, batches, schedules, saved results and usage workflows through Caelario.

Source: https://github.com/chunkydonut21/caelario-mcp

The official local server gives Codex and Claude Code typed tools for capability discovery, immediate and durable extraction, result pagination, and credit estimates.

Install for Codex

bash
npx -y caelario-mcp install codex

Install for Claude Code

bash
npx -y caelario-mcp install claude

The installer prompts for a Caelario workspace API key and registers the stdio server at user scope. It does not ask for Instagram, TikTok, YouTube, X, or Facebook credentials; provider identities are managed by Caelario.

After installation, restart the client and ask:

Use Caelario to get Nike's Instagram profile and 10 recent posts.

The package requires Node.js 20 or newer. See https://caelario.com/mcp-server for security, billing, and tool details.

The connector is released under the MIT License. Caelario's hosted API, provider implementations, acquisition infrastructure, and managed identities are separate proprietary services and are not included in this package.

The npm artifact also includes official MCP Registry metadata under io.github.chunkydonut21/caelario. npm publication must happen before the separate Registry publication step.

来源:README.md,提交 b3d6a14

工具

21

discover_platforms

List every Caelario platform, operation, limit, and price.

无需输入参数。

get_operation_guide

Explain and validate one operation before spending credits.

输入参数platform必填operation必填

get_account

Return workspace identity and extraction limits.

无需输入参数。

get_credit_usage

Return available, reserved, promotional, purchased, and spent credits.

无需输入参数。

estimate_credits

Estimate the maximum reservation for one or many targets before starting work.

输入参数limitplatform必填operation必填target_countinclude_posts

extract_now

Run one bounded target synchronously and return a model-safe result page.

输入参数limitcursortarget必填platform必填operation必填text_limittext_offsetresult_limitinclude_postsresult_offsetidempotency_key

create_extraction

Create one durable extraction job for work that may continue in the background.

输入参数limitcursortarget必填platform必填operation必填include_postsidempotency_key

list_extractions

List recent durable extraction jobs without embedding their large results.

输入参数limit

get_extraction

Get one job's progress, billing, and error metadata.

输入参数job_id必填

get_extraction_result

Page through a completed job's structured result without flooding model context.

输入参数limitjob_id必填offsettext_limittext_offset

wait_for_extraction

Wait briefly for a durable job to become terminal without returning its large result.

输入参数job_id必填timeout_secondspoll_interval_seconds

cancel_extraction

Cancel a queued job or request cancellation of a running job.

输入参数job_id必填

create_batch

Run one operation for up to 100 targets as an expandable durable batch.

输入参数limittargets必填platform必填operation必填include_postsidempotency_key

list_batches

List recent batches and optionally their child jobs.

输入参数limitinclude_jobs

get_batch

Get aggregate status and child jobs for one batch.

输入参数batch_id必填

cancel_batch

Cancel queued batch children and request cancellation of running children.

输入参数batch_id必填

list_schedules

List active and paused extraction schedules.

输入参数limit

get_schedule

Get one schedule and its recent run history.

输入参数schedule_id必填

create_schedule

Create a timezone-aware once, daily, weekly, or monthly extraction schedule.

输入参数name必填limittargets必填platform必填timezone必填frequency必填operation必填starts_at必填include_posts

update_schedule

Replace a schedule definition while preserving its run history.

输入参数name必填limittargets必填platform必填timezone必填frequency必填operation必填starts_at必填schedule_id必填include_posts

manage_schedule

Pause, resume, run now, or delete an extraction schedule.

输入参数action必填schedule_id必填

版本历史

1
  1. v0.1.5Sep 16, 2026

评分

已满足适用检查项的 50%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通25 分

    请保持端点能够响应 MCP 握手;SourceWeft 每 30 天探测一次。

  • 已声明认证方式10 分

    请在注册中心条目中声明凭据请求头,或保持端点能够响应探测,以便判定其认证方式。

  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分