Survey Light

com.mobopinions由 mobopinions.com 发布官方发布者v1.0.0更新于 Oct 24, 2026

Build, target, launch and analyse consumer surveys from your AI assistant (Survey Light, 71 tools).

已验证Streamable HTTP可网页运行数据与分析商业与电商

概览

AI 生成的概览

MobOpinions Survey Light 的远程 MCP 服务器，让助手在工作区内创建、定向、投放并分析付费消费者问卷。

可以做什么
  • 管理账户与工作区：查看资料，列出、创建、重命名、切换工作区，邀请或移除成员
  • 构建问卷：创建草稿，按问题或模板生成，克隆、翻译、预览、删除
  • 编辑问卷内容：增删改排序题目与选项，设置跳转逻辑，解析问卷文档
  • 设置受众与配额：人口与地域定向、配额单元、条件与目标完成数
  • 投放与执行：完成构建、查看投放摘要与报价、启动、暂停、恢复、关闭
  • 查看结果：汇总、交叉表、AI 摘要、洞察报告、CSV 与 Excel 导出、完成记录、钱包余额与充值结账
适用场景
适合希望由助手起草消费者问卷、设置定向与配额、在 MobOpinions 样本库投放并直接获取结果或导出文件的场景。适用于已在 Survey Light 上开展市场研究、并愿意按完成份数付费的团队。
运行要求
远程 Streamable HTTP 端点，无需安装或 API 密钥。需用现有 surveys.mobopinions.com 邮箱和密码完成 OAuth 2.1 登录，令牌有效期 30 天并自动刷新。工具以登录用户权限在所选工作区内运行。实际投放需要工作区钱包有余额。
安装前请注意
启动问卷会从钱包预留整个项目金额，关闭时按已接受的完成份数扣费；投放中提高目标会预留更多资金，充值会生成 Stripe 结账链接。许多工具会写入或删除数据，包括不可逆的题目与问卷删除、移除成员和撤销邀请。助手读取结果和受访者级导出时，数据会离开工作区。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Survey Light，将其添加到工作区。
  2. 连接 所需的 凭据。它们会被私密存储，绝不会在公开页面上展示。
  3. 为需要使用其工具的对话启用该服务。

Web executable，通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体，或 SourceWeft 对话中。智能体会读取该服务器的安装指南，先向你展示发布者、端点和登录方式，经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/com.mobopinions%2Fsurvey-light/install.md，并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前，先告诉我它的发布者、端点和登录方式，等我确认。不要把 API 密钥写进命令或我会提交的文件；问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。 客户端会在首次连接时完成 OAuth 登录。

{
  "mcpServers": {
    "survey-light": {
      "type": "http",
      "url": "https://surveys.mobopinions.com/light-api/mcp"
    }
  }
}

README

MobOpinions Survey Light — MCP server

Run consumer surveys from your AI assistant. Every action in Survey Light (MobOpinions' self-service survey platform) is exposed as a Model Context Protocol tool, so Claude, ChatGPT, Cursor, VS Code or any MCP client can create a survey, write the questions, target a real consumer audience through the MobOpinions panel network, launch it and read the results — inside your own MobOpinions workspace.

Registry name: com.mobopinions/survey-light · Remote server (no install) · License: MIT

Endpointhttps://surveys.mobopinions.com/light-api/mcp
TransportStreamable HTTP (MCP 2025-06-18), stateless JSON responses
AuthOAuth 2.1 — RFC 9728/8414 discovery, RFC 7591 dynamic registration, PKCE S256, refresh tokens
Sign-inyour existing surveys.mobopinions.com email + password
Tools71

Connect

Claude Code

bash
claude mcp add --transport http surveys https://surveys.mobopinions.com/light-api/mcp# then in a session: /mcp → surveys → Authenticate (browser login)

Claude.ai / Claude Desktop — Settings → Connectors → Add custom connector → URL https://surveys.mobopinions.com/light-api/mcp → Connect → sign in.

ChatGPT — Settings → Connectors → Create (developer mode) → MCP server URL https://surveys.mobopinions.com/light-api/mcp, authentication OAuth → sign in.

Cursor / VS Code / any Streamable-HTTP + OAuth client — { "mcpServers": { "surveys": { "url": "https://surveys.mobopinions.com/light-api/mcp" } } }

There is no API key: the client discovers the authorization server, registers itself, opens the login page and stores the token. Tokens last 30 days and refresh automatically. Every tool runs with exactly the permissions of the signed-in user, in the workspace they select.

What you can do

Ask your assistant things like:

  • "Create a 6-question survey about oat-milk buying habits for 300 UK adults aged 25–44 and launch it."
  • "Add a screener so only people who bought running shoes in the last 6 months qualify."
  • "How many completes do we have on the Berlin survey, and what's the top answer to question 3?"
  • "Export the results as an Excel file and pause the survey."

Tools (71)

Account & workspaces (12)

ToolWhat it does
get_profileWho am I: the linked account (user id, email, display name), the workspaces
get_accountAccount card: full name, email, Survey Light workspaces (id, name, role,
update_accountSet the account's full name and optionally rename the ACTIVE workspace.
list_workspacesMy Survey Light workspaces plus the id of the one this session is switched to.
switch_workspaceMake a workspace the active one for this session. Surveys, wallet and
create_workspaceCreate a new Survey Light workspace (you become its owner) and switch to it.
rename_workspaceRename a workspace you own.
list_workspace_membersMembers (user id, email, role) and pending invitations of a workspace you own.
invite_workspace_memberEmail an invitation to join a workspace you own (link valid 7 days).
remove_workspace_memberRemove a user from a workspace you own. Confirm with the user first.
cancel_workspace_invitationRevoke a pending workspace invitation.
accept_workspace_invitationRedeem a workspace invitation token (from the invitation email) for the

Surveys (16)

ToolWhat it does
get_survey_optionsReference lists for creating a survey: countries available for fielding
list_surveysSurveys in the active workspace (newest first, max 300) with status
create_surveyCreate an empty draft survey in the active workspace. Then add questions
suggest_survey_topicAI-pick the best topic category (from get_survey_options.categories) for a research question.
create_survey_from_questionBuild a complete survey (questions, screeners, skip logic, images) from ONE
create_survey_from_templateCreate a new draft by copying a template survey (a survey marked with
clone_surveyDeep-copy a survey (questions, answers, targeting, skip logic) into a new draft.
translate_surveyClone a survey once per language and machine-translate every question and
set_survey_template_flagMark (or unmark) a survey as a reusable workspace template.
preview_surveyGet a respondent-view preview link for a survey and mark it as tested.
pause_surveyPause a live survey (stops collecting; funds stay reserved). Resume with resume_survey.
resume_surveyResume a paused survey that was previously launched.
delete_surveyPermanently delete a draft or closed survey with no completes. Live/paused
get_survey_statusCheap liveness snapshot for a survey in field: status, is_live, accepted/
launch_surveyTake a draft live: reserves the full project value from the wallet as a hold
close_surveyClose a running survey for good: stops fielding, charges accepted completes

Questions & answers (13)

ToolWhat it does
list_questionsThe questionnaire: every question with type, text, answer options, matrix
add_questionAppend a question with its answer options. single_select / multi_select /
update_questionEdit an existing question's text, mandatory flag, answer randomisation,
delete_questionDelete a question and its answers/statements. Irreversible.
add_answer_optionAdd one answer option to a single_select / multi_select / ranking / matrix question.
update_answer_optionChange an answer option's text, flags, image or position.
remove_answer_optionDelete one answer option from a question.
reorder_answer_optionsSet the display order of a question's answers by listing their codes in order.
reorder_questionsSet the order of the editable questions by listing their qids. Locked
set_question_routingSkip logic for a single_select question: per answer code, jump to a later
parse_questionnaire_documentAI-parse a questionnaire document (PDF, DOCX, XLSX or CSV, max 20 MB) fetched
get_complete_answersFull question-by-question answers of one complete (by location_id from list_completes).
get_close_previewWhat closing would settle: accepted/rejected counts, held amount, charge and release-back.

Audience & targeting (14)

ToolWhat it does
get_targetingCurrent audience: country, gender, age ranges, employment, education,
set_targetingReplace the survey's demographic targeting (each list you pass REPLACES that
convert_location_textTurn a free-text location description (e.g. 'Texas and Florida', 'Greater
set_location_targetingApply location targeting: EITHER an admin-area selection (assoc_type + values,
list_quotasQuota cells of a survey with limit, filled, remaining and their conditions
create_quotaCap how many completes a group may fill. Demographic cell: quota_type 'gender'
update_quotaChange a quota's limit (any cell) or name/descriptor (manual cells only).
delete_quotaDelete a manual quota cell (auto cells follow targeting — freeze them instead).
freeze_auto_quotaFreeze an auto-generated (targeting-derived) quota so it never fills and is
add_quota_conditionAdd a (question, answer code) condition to a manual quota. A condition on a
update_quota_conditionChange the question and/or answer code of a quota condition.
remove_quota_conditionRemove one condition from a manual quota.
split_quota_limitsSpread the completes target across cells — e.g. scope 'gender' = 50/50
set_target_completesChange the target number of completes. In field, raising it reserves the

Launch & field (2)

ToolWhat it does
finish_builderCall after editing questions: recomputes LOI from the final question count
get_launch_summaryPre-launch check: final quote (question count, LOI, CPI, total), whether the

Results, insights & export (6)

ToolWhat it does
get_resultsPer-question aggregates over accepted completes: counts and percentages for
export_results_csvRaw respondent-level CSV: one row per complete (accepted and rejected), one column per question.
export_results_xlsxExcel workbook of every complete: 'Responses' (labels), 'Codes' (stored
get_ai_results_summaryAI-written digest of the results: headline, key points and watch-outs
get_insights_reportThe full Insights report: methodology, banner columns, per-question findings
run_insights_functionAdvanced Insights operations by name: AI narrative (insights_ai_summary with

Wallet & billing (2)

ToolWhat it does
get_walletWallet balance (total_credits, pre-formatted with currency symbol) and the
create_wallet_topup_checkoutCreate a Stripe Checkout link to add funds to the wallet. Returns checkout_url —

Other (6)

ToolWhat it does
get_price_quotePrice a hypothetical survey without creating anything: incidence estimate,
list_piping_sourcesQuestions whose answer can be piped into later text with the {{Q}} token
upload_mediaUpload an image or video from a public URL to use in question text or as an
list_completesEvery complete (accepted and rejected) with timestamps, duration and rate,
reject_completeReject a low-quality complete before close (excluded from billing, slot
get_crosstabCross-tabulate one question (rows) by up to 4 banner questions (columns) with

Pricing

Building, editing and analysing surveys is free. Fieldwork (real respondents) is paid per completed response from the workspace wallet, at the price shown before you launch. Using the MCP server adds nothing on top.

Support

来源：README.md，提交 53059e4

工具

0
该服务器需要登录，因此 SourceWeft 无法在没有账号的情况下读取其工具，也未收录。已连接账号所能看到的内容可能因账号及其权限而异。

版本历史

1
  1. v1.0.0最新Sep 18, 2026

评分

已满足适用检查项的 85%

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 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 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

其他 Mobopinions 集成

Adtest Mcp

ai.adtest

AdTest.AI 的本地 MCP 服务器，让助手按 13 个维度为图片、视频或文字广告打分并返回详细报告。

STDIO仅桌面
商业与电商尚未验证Oct 24, 2026