原型 · 预览版:收录官方注册中心里的全部服务器,每 30 分钟同步一次。评价和账户功能将在后续阶段推出。
PopNx MCP为使用 AI 的人准备的 MCP 服务器 · 用大白话讲清楚

flconsole — Upwork jobs

作者:droopboop版本 1.0.0i★ –i在云端运行
注册名称com.flconsole/upwork-jobs

安装

概览

作者提供的说明(英文原文)

Free hosted MCP server: new Upwork jobs for your agent and Telegram alerts within a minute.

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。

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: 聊天与邮件

安装

选择你使用的 AI 应用,然后复制设置。

安装之前

你需要

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:

  1. 打开 Settings → Connectors
  2. 点击 Add custom connector
  3. 粘贴下面的地址,然后点击 Add
  4. 如果要求登录,请登录作者的服务
SettingsConnectors1+ Add2https://…3
https://mcp.lite.flconsole.com/mcp

在终端中运行这条命令

claude mcp add --transport http upwork-jobs https://mcp.lite.flconsole.com/mcp

在 Cursor 中打开

或者自己复制设置

粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json

{
  "mcpServers": {
    "upwork-jobs": {
      "url": "https://mcp.lite.flconsole.com/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。

在 VS Code 中打开

或者自己复制设置

粘贴到项目中的 .vscode/mcp.json

{
  "servers": {
    "upwork-jobs": {
      "type": "http",
      "url": "https://mcp.lite.flconsole.com/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:

  1. 打开网页版 ChatGPT → Settings → Connectors (Apps)
  2. 点击 Create 或 Add(部分套餐需要先在 Advanced 中开启 Developer mode)
  3. 粘贴下面的地址,起个名字,然后保存
  4. 如果要求登录,请登录作者的服务
SettingsConnectors1+ Create2https://…3
https://mcp.lite.flconsole.com/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

我们还没有读取这个服务器的工具列表;我们会轮流检查各个端点。

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7日

  • ✓ 通过

    作者与仓库一致

    注册名称已通过域名 flconsole.com 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

  • … 尚未检查

    端点有响应

    尚未检查;我们会轮流检查各个端点

  • … 尚未检查

    使用 OAuth 登录

    尚未检查;我们会轮流检查各个端点

  • … 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“聊天与邮件”中的其他服务器比较i

  • 星标数多于此分类中其他 1785 个服务器的 0%
  • 发布到注册中心的时间比此分类中其他 1785 个服务器的 45% 更新
  • 基础检查结果等于或优于此分类中其他 1785 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

flconsole — Upwork jobs 安全吗?

它在 1 项适用的基础检查中通过了 1 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(1.0.0)于32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:Vexa, emailmd, MCP WhatsApp Cloud

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:★ 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

Vexa

作者:Vexa-ai

Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.

在云端运行★ 2,949Apache-2.02个月前更新
安装

emailmd

作者:anypost

Render markdown into email-safe HTML, lint drafts for deliverability problems, and preview emails.

电脑或云端★ 1,385MIT32天前更新
安装

MCP WhatsApp Cloud

作者:codespar

MCP server for WhatsApp Cloud API (Meta) — messages, media, and templates via Graph API

在你的电脑上运行★ 277MIT5个月前更新
安装

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。

flconsole — free Upwork MCP server

New Upwork jobs for Claude Code, Codex, Cursor and the Claude app.

Your agent finds new jobs, reads them and shows you the ones that fit.

Website · Docs · Telegram bot


flconsole is a free service that watches new Upwork jobs and delivers matching ones to your Telegram or to your AI agent over MCP.

This repository holds ready-to-paste configs and a short guide. The server is hosted — there is nothing to install or run.

https://mcp.lite.flconsole.com/mcp
  • Free. No card, no API key, no paid plan.
  • One feed, for your agent and your phone. Your agent saves searches as feeds on the server; turn on alerts for any of them with one command and the Telegram bot sends matching jobs within a minute of posting.
  • No access to your Upwork account. flconsole never asks for your Upwork login.

Contents

  • Connect — Claude Code, Codex, Claude app, Cursor, other clients
  • Sign-in
  • Three ways to work
  • Tools
  • Prompts to start with
  • Skill for Claude Code
  • Feed tuner: keep your alerts accurate
  • No terminal? Install in Claude app
  • With Upwork's own MCP
  • Limits
  • Links

Connect

Claude Code
claude mcp add --transport http flconsole https://mcp.lite.flconsole.com/mcp

Add --scope user to use it in all your projects. Then run /mcp → flconsole → Authenticate. Guide: flconsole in Claude Code · file: configs/claude-code.sh

Codex
codex mcp add flconsole --url https://mcp.lite.flconsole.com/mcp

codex mcp login flconsole

The second command opens the sign-in page in your browser. Guide: flconsole in Codex · file: configs/codex.sh

Claude app

Web, desktop and mobile: Customize → Connectors → + → Add custom connector, paste the server URL:

https://mcp.lite.flconsole.com/mcp

Guide: flconsole in the Claude app

Cursor

Add to ~/.cursor/mcp.json, then sign in when Cursor asks:

{

"mcpServers": { "flconsole": { "url": "https://mcp.lite.flconsole.com/mcp" } } }

Guide: flconsole in Cursor · file: configs/cursor.mcp.json

Other clients

Any MCP client with Streamable HTTP works. If it supports sign-in (OAuth), add the server URL and sign in as below. If it doesn't, use a manual token: in the Telegram bot send /mcp → Manual token (other apps), then send it as Authorization: Bearer . Examples: configs/manual-token.claude-code.sh, configs/manual-token.codex.toml.

Sign-in

Through Telegram — no password, no API key:

  1. Your client opens a sign-in page in the browser.
  2. Tap Continue in Telegram (or scan the QR code). The flconsole bot opens; if you're new, tap Start.
  3. The bot names the app that wants access. Tap Allow and get a 4-digit code.
  4. Enter the code on the sign-in page. Done.

The sign-in link works for 10 minutes. Disconnect any app later in the bot: /mcp → Disconnect.

Three ways to work

WayWhat your agent doesWhat is stored
1. Search onlyfind_jobs with a search query and filters in every callNothing
2. Feeds for your agent + memoryStarts each session with get_agent_context, keeps up to 5 feeds with alerts off (search, job and client filters, your AI filter in plain words, its own notes) and its notes about you. Start any agent anywhere and it picks up where the last one stoppedFeeds and the agent's notes, on flconsole's server
3. Telegram alertsTurns on alerts for a feed with one command (set_feed_alerts): the Telegram bot sends you every new matching job, filtered by the same AI filterThe same feed, now with alerts on

A feed with alerts off is for your agent only: nothing goes to Telegram and nothing runs on the server — your agent checks it with find_jobs and applies your AI filter itself. You see both kinds in the bot's /feeds, where you can turn alerts on or off, pause or delete a feed.

Tools

ToolWhat it does
get_agent_contextStart of a session: how-to, the agent's notes about you and all your feeds
find_jobsNew jobs by a feed, all feeds or any filters
get_job_detailsFull data for the jobs you picked
list_feedsYour feeds: filters, AI filter, alerts on or off, paused
create_feedSaves a feed (alerts off unless you ask); needs your approval
update_feedChanges a feed's search, filters, AI filter or the agent's notes
set_feed_alertsTurns Telegram alerts for a feed on or off; needs your approval
set_feed_pausedPauses or resumes a feed
delete_feedDeletes a feed; needs your approval
save_agent_instructionsSaves the agent's notes about you: preferences, conclusions
preview_filtersHow many jobs matched over the last 7 days: total, ~per day and per UTC day, with samples
parse_upwork_urlTurns an Upwork search link into filters
list_categoriesUpwork categories and subcategories
find_locationsCountries and regions by name
get_usageWhat's left of your limits
get_skillStep-by-step instructions for a task: job scout, Telegram setup, feed tuner, AI filter, ratings
rate_job👍/👎 a job for you, only when you say so
tuner_checkinFeed tuner: which feeds are due for a check now
get_feed_reportFeed tuner: jobs sent and opened, your 👍/👎, what the AI filter skipped, recent changes
set_feed_tunerTurns the feed tuner on or off for a feed, with your target jobs a day; needs your approval
log_tuner_runFeed tuner: records what was checked and changed; you get the change in Telegram with Undo

Arguments, filters and job fields: MCP reference.

Prompts to start with

Find new React Native jobs: hourly from $40, clients from the US or Canada with verified payment. Show only the ones about fintech.
Check new jobs in my flconsole feeds and show only the ones that fit.
Save this Upwork search link as a feed: <link>. Show me the filters before saving.
Turn on Telegram alerts for my React Native feed.

On a schedule in Claude Code:

/loop 10m check new jobs in my feeds and show only relevant ones

Skill for Claude Code

skills/upwork-job-hunter turns your agent into a job scout. No feed needed: the agent searches new jobs with your query and filters and reads every job itself.

  • Picks up where you left off. Every session starts with get_agent_context: your feeds and what the agent learned about you are kept on the server, so any agent on any device continues the same work. Prefer a local file? Say so, and it keeps upwork-job-hunter.md instead.
  • Setup in 3 steps, like the onboarding of the Telegram bot: what to search, job and client filters, and your own AI criteria in plain words. The agent counts matching jobs at every step and suggests how to sharpen the search.
  • Your AI filter, applied by your agent. It tests your criteria on real jobs, shows what passed and why, and you adjust them.
  • Shortlist with reasons on every check, and proposal drafts for the jobs you pick. Nothing is sent for you.
  • Tuning with numbers. Too many jobs, too few or not the right ones — the agent shows the counts and offers concrete changes to the search, filters and criteria.
  • Telegram alerts when you want them: one command turns them on for the same feed, with the same AI filter.

Install it with the skills CLI (add -g to use it in all your projects):

npx skills add droopboop/flconsole-upwork-mcp

Or copy it into ~/.claude/skills/:

mkdir -p ~/.claude/skills/upwork-job-hunter && curl -fsSL https://raw.githubusercontent.com/droopboop/flconsole-upwork-mcp/main/skills/upwork-job-hunter/SKILL.md -o ~/.claude/skills/upwork-job-hunter/SKILL.md

Then ask Claude Code:

Set up my Upwork job search.

Feed tuner: keep your alerts accurate

skills/feed-tuner lets your agent keep your Telegram job alerts accurate on its own.

  • Set up once. The agent asks a few questions in one message — jobs a day, work you never take, budget and client limits — checks your feed on the last 7 days and shows a short plan.
  • Checks on a schedule. An hourly task in Claude runs the check; the server decides when a feed is due (often for a new feed, rarely for a stable one).
  • Looks at what matters: the jobs against what you asked for, how many arrive against your target, what the AI filter skipped, and your 👍/👎.
  • Small, explained changes. At most one set of related edits per check, to the search, filters or AI filter. Every change comes to Telegram with the reason and an Undo button.
  • Your edits win. It never deletes or pauses a feed and never turns alerts off.

In Claude (web, desktop) with the flconsole connector, just say:

Set up the flconsole feed tuner for me.

The tuner rules are served by flconsole (get_skill), so no install is needed. For Claude Code, add the skill like the one above:

mkdir -p ~/.claude/skills/feed-tuner && curl -fsSL https://raw.githubusercontent.com/droopboop/flconsole-upwork-mcp/main/skills/feed-tuner/SKILL.md -o ~/.claude/skills/feed-tuner/SKILL.md

More: flconsole.com/agent-feed-tuner

No terminal? Install in Claude app

Works in the Claude app on the web and on desktop.

  1. Open Customize → Connectors, click + → Add custom connector and paste https://mcp.lite.flconsole.com/mcp. Click Connect and sign in through Telegram (see Sign-in).
  2. Quickest: paste a prompt. Copy the [ready prompt](https://flconsole.com/how-to/mcp?utm_source=github&utm_medium=referral&utm_

只显示了开头部分;其余内容请在仓库中阅读

问 PopNx