In short: the API reads and writes your data from your systems. The MCP server lets an AI assistant like Claude query your feedback. The display widget publishes your ratings and reviews on your website.
The Feedier REST API
Menu → Settings → Advanced settings → API (admins only)
API documentation: https://developers.feedier.com/
Element | Use |
Documentation URL | The complete endpoint reference |
Endpoint URL | Your workspace's API base address |
Public key | A non-sensitive identifier, used by widgets |
Private key | Your access token for the public API ( |
Authenticate each call with the private key:
Authorization: Bearer <your_private_key>
The private key must stay secret. It's reserved for server-to-server exchanges and must never appear in a browser-side or mobile app.
Use case | How |
Send a survey after each purchase | Trigger an existing campaign with the customer's attributes. See Distribute a survey. |
Extract feedback to your data warehouse | Read feedback, filtered with FQL |
Manage unsubscribes |
|
Sensitive attributes remain readable via the API. A private key gives access to all the workspace's data: to isolate data, use separate workspaces.
The Feedier MCP server
Feedier exposes an MCP (Model Context Protocol) server. A compatible assistant, such as Claude or Copilot, can then measure your KPIs, explore your topics, and generate reports and action plans, within the limits of your permissions.
Tools | Ping, Scope, Measure, Explore, Discover, Report, GenerateActionPlan |
Resources | Topics, Attributes, Segments, Questions, KPIs |
Example: connect Claude
In Claude, open the connectors settings and add a custom connector.
Name:
Feedier. URL:https://bx.feedier.com/mcp(orhttps://<your-domain>/mcpwith a custom domain).Sign in with your Feedier account and approve access.
Ask your questions: "Measure the NPS for the last 30 days", "Explore the main topics from last week", "Generate an action plan from detractors".
The assistant only sees what your connected Feedier account can see: your role, your teams and attribute visibility all apply.
The review display widget
Display an average rating and customer reviews on your website, with no complex development. There are two modes: compact (a rating) and full (a list of review cards).
Add the script to your website's header, with your public key:
<script async defer src="https://display-widget.feedier.com/widget.js?api_key={PUBLIC_KEY}"></script>Place a block where you want it:
<div data-feedier-widget data-layout="compact" data-primary-color="#cc0d2b" data-kpi='ratings("survey.id", "=", "123")' data-score-max="5"></div>Parameter | Required | Role |
| ✅ |
|
| ✅ | The KPI displayed, written as a KPI formula |
| ✅ | The scale maximum (5 for a rating, 10 for NPS) |
| The color, in hexadecimal | |
| Only displays the widget above this number of feedback | |
| Only displays the widget if the score exceeds this threshold | |
| An FQL filter on the feedback displayed | |
| The language ( | |
| The element to point the click to, for example the full widget | |
| HTML text below the title | |
| The title, rating and verbatim of each card |
Tip: combine data-min-count and data-threshold to only show the rating when it's reliable and flattering, for example at least 200 reviews and an average above 3.5.
Workspace webhook
A webhook can be triggered when a new user joins Feedier, to add them to your own tools. For feedback, use an automation instead.
FAQ
Where do I find a survey or question ID?
In the URL of the corresponding page, or in the Options → Survey management tab in Feedier Surveys.
Can I regenerate the private key?
Contact support: all calls that use the old key will need to be updated.
Last reviewed: November 11, 2026, version 4.23.5.


