Survicate overview

Survicate is a customer feedback platform: targeted surveys on websites and web apps, in mobile apps, by email and shareable link, and inside Intercom Messenger and Braze, plus Research Hub, which gathers feedback from surveys and other sources and analyzes it with AI. This page defines the concepts the developer documentation builds on and shows how the SDKs, the Data Export API, Webhooks and Survicate MCP fit together.

Surveys are created, targeted, launched and analyzed in the Survicate panel at panel.survicate.com. Developers come in where Survicate meets your own code and systems: installing the SDKs, identifying respondents, reporting events, and getting responses out. Using the panel itself (building surveys, targeting, integrations, billing) is documented in the Help Center.

Survey types

A survey has exactly one type, chosen when it is created. The type decides where the survey is shown, which question types are available and what, if anything, a developer has to install. It cannot be changed later, except between Website or in-product and Braze in-app message.

Question types: welcome message, single answer selection, multiple answer selection, text answer (up to 1,000 characters), rating scale (2 to 10 points, used for CSAT), smiley scale (3 or 5 points), NPS (fixed 0 to 10), dropdown list, matrix (up to 13 rows and 7 columns), contact form, ranking, date, and a thank-you screen with an optional call to action. Not every type is available in every survey type; the Help Center has the full matrix, and Webhooks: Answer types shows how each answer arrives in your systems.

How a survey is delivered

1. Your page or app loads the SDK with the workspace key. The SDK assigns the visitor a UUID, kept in the browser's Local Storage or on the device, so the same person is recognized on later visits.
2. Your code can pass attributes about the visitor (such as user_id or plan) and report events and screens. Everything else about who sees which survey is configured in the panel and takes effect without a deployment.
3. The SDK works out which surveys the visitor is eligible for. For a Visitors audience the conditions are evaluated in the browser or app from data stored there. For Users and Manual audiences Survicate evaluates them on its servers, keyed by user_id, and the result is cached for about five minutes.
4. When the survey's trigger fires (page or screen load, a delay, exit intent, or an event), the SDK renders the survey. Each answer is sent to Survicate as soon as it is given, so partial responses are stored as well.
5. Survicate records the response, notifies the survey's integrations and webhooks, and shows it in the survey's Analyze tab, from where it also feeds Boards and Research Hub.

Concepts

• Organization and workspace. An organization is a company's account. It holds one or more workspaces, and each workspace has its own surveys, integrations, teammate permissions and workspace key.
• Workspace key. The identifier the tracking code and the Mobile SDKs use to load the workspace's surveys, as in https://survey.survicate.com/workspaces/<workspace key>/web_surveys.js. It is embedded in your pages and apps, so it is not a secret. It is shown under Settings → Organization → Access Keys in the panel.
• Access keys. Besides the workspace key, a workspace has an API key for the Data Export API and an optional secret key used to sign user identities for logged-in user targeting. Webhooks use their own verification token and signing secret. Getting started lists where each one lives and who can see it.
• Teammates and roles. Four roles: Guest (view results), User (create and edit surveys, connect integrations, export results), Admin (also invites teammates and manages folder access) and Owner (also adds workspaces, deletes teammates and manages the subscription). Choosing roles per workspace is available on some plans; otherwise every teammate is a User. Access to Research Hub is granted separately from the role. Survey folders can be open to everyone or limited to chosen teammates and teams.
• Survey, question and answer. A survey is identified by a 16-character hexadecimal ID that appears in the panel's address bar when the survey is open, for example f658c90277553239. Questions (called points in the JavaScript SDK) and answer choices have numeric IDs. A survey is enabled or disabled, can be scheduled to start and stop, can cap its responses, and can be archived.
• Respondent and response. A respondent, also called a visitor, is one person identified by a UUID plus any attributes you passed. A response is one respondent's submission to one survey; complete and partial submissions are both stored and counted, test responses included, and any response can be deleted in the Analyze tab. Submitting only the welcome screen is shown in the analysis but not counted.
• Attributes (user traits). Name-value pairs describing a respondent, set from your code, an integration, a CSV upload or the survey URL. Values are strings, numbers, booleans or dates. Attributes identify respondents, define audiences, filter results and can be recalled in question text. Four names are special: user_id marks the respondent as logged in and enables server-side targeting, and first_name, last_name and email make a response identified instead of anonymous. Names are case-sensitive, names and values are limited to 255 characters, and a workspace can hold up to 2,000 distinct attribute names. Response attributes belong to a single response rather than the respondent, for example the order a review refers to.
• Events. Named actions your code reports with invokeEvent, optionally with string properties. In the panel an event can trigger a survey the moment it happens (counted in the browser or app) or filter a Users audience by the user's server-side event history, which follows the user across devices.
• Audience. The group of respondents a survey targets; All visitors is the default and All logged-in users is everyone with a user_id. A Visitors audience is evaluated in the browser or app from attributes stored there. Users and Manual (CSV upload) audiences are evaluated on Survicate's servers by user_id, so the same person is recognized across browsers and devices and cannot answer twice. A survey uses either browser-side or server-side targeting, not both.
• Targeting. Everything else that decides when a survey shows, set in the panel: pages or screens (URL rules, regular expressions), triggers (on load, after a delay, on exit intent, on an event), device, browser language, visit frequency and depth, sampling, how often a respondent may answer again, schedule and response caps.
• Integrations. Native connections that send responses to other tools, sync respondent data, or pull feedback into Research Hub: Slack, Microsoft Teams, Google Sheets, HubSpot, Intercom, Salesforce, Segment, RudderStack, Amplitude, Mixpanel, Google Analytics, Google BigQuery, Zendesk, Jira, Productboard, Zapier and others. Availability depends on the plan and the survey type; the Help Center's Connect integrations collection covers each one.
• Research Hub. Survicate's research and analysis product. It collects feedback pieces from surveys, review sites (App Store, Google Play, G2, Google Reviews), support tools (Intercom, Zendesk, Front, LiveChat), call and meeting transcripts (Gong, Google Meet, tl;dv, uploaded recordings), Slack, email and uploaded files. A Research Project runs an AI analysis of chosen sources against a research question and produces a report whose conclusions link to the source quotes, plus Insights, the recurring themes found in the data. The Research Agent answers questions inside a project.
• Boards. Dashboards within a workspace that combine widgets from several surveys and from Research Hub: single question, combined score, and Research Hub breakdowns by sentiment, insight or attribute.
• Survicate AI. The AI features across the product: creating a survey from a description, AI follow-up questions to text answers, Insights that summarize text answers, Research Hub's analysis and Research Agent, and Survicate MCP. Text analysis uses OpenAI's Enterprise API, which does not train models on customer data.

Developer surfaces

Getting data in

• Respondent attributes and events, from the SDKs, Google Tag Manager, Segment, RudderStack, or URL parameters on a survey link.
• Respondent profiles, from integrations such as HubSpot, Intercom or Clearbit.
• Audiences, from a CSV upload of user IDs.
• Feedback from other sources into Research Hub, through integrations, transcribed recordings and file uploads.

Getting data out

• As it happens, per answer or per response: Webhooks, or a native integration (Slack, Google Sheets, HubSpot, Amplitude, Mixpanel and others).
• In bulk or as a backfill: the Data Export API, or the Google BigQuery and Fivetran integrations for a warehouse.
• On demand: CSV and XLSX exports and PDF or PNG charts from the Analyze tab.
• In conversation: Survicate MCP, which answers with citations to the underlying responses.

Limits and defaults

Plans

Survicate has a free plan (25 responses a month, up to 3 teammates, no data export) and paid plans: Starter (monthly), Growth, Pro and Enterprise (yearly). A new account starts with a 10-day trial of the Growth features. Installing the tracking code or a Mobile SDK, identifying respondents with setVisitorTraits and URL targeting work on every plan; the developer features below are tied to plans. The pricing page and its Compare plans table are authoritative; this list reflects them as of September 2026.

• Attribute-based audiences, recurring surveys and multilingual surveys: Growth and above.
• Event-based targeting, JavaScript targeting (the SDK methods other than setVisitorTraits, and event listeners), sampling, A/B testing and custom CSS: Pro and Enterprise.
• Webhooks and the Data Export API: Pro and Enterprise.
• Survicate MCP: all paid plans.
• SAML SSO, access logs, HIPAA and DORA terms: Enterprise.

Security and compliance

• Hosted on AWS in Ireland. Backups are encrypted and stored in at least three regions in Ireland.
• Data is encrypted in transit (TLS 1.2 or higher only) and at rest (AWS KMS for databases and file stores).
• Product development and maintenance are ISO 27001 certified; the infrastructure is covered by AWS's ISO 27001 and SOC 2 programs. Penetration tests run at least yearly, and there is a bug bounty program. Security issues go to security@survicate.com.
• GDPR: a data processing agreement is available, and the Data Export API's personal data endpoints retrieve or delete one person's data across services. HIPAA (with a BAA) and a DORA addendum are available on the Enterprise plan; HIPAA-regulated workspaces cannot be connected to AI assistants through MCP.
• Account security: two-factor authentication for every account, SAML 2.0 single sign-on (Okta, Microsoft Entra ID, OneLogin and other SAML providers) on Enterprise.
• Respondent data: Survicate sets no cookies on respondents' devices. The web SDK keeps the respondent UUID, survey status, tags, queued answers and attributes in Local Storage, or only in Session Storage when disableSensitiveDataPersistence is set. IP addresses are processed for CDN delivery and abuse protection and are not stored.
• Accessibility: surveys work with screen readers (VoiceOver, NVDA, TalkBack), full keyboard navigation and zoom up to 400%, with the goal of WCAG 2.1 A and AA and Section 508 conformance.
• Service status: status.survicate.com.

Docs for AI agents

Every page of this site is also published as plain Markdown for AI agents and tools: append .md to a page URL (for example https://developers.survicate.com/getting-started.md; this page is https://developers.survicate.com/index.md), fetch llms.txt for an index of all pages with one-line descriptions, or llms-full.txt for the whole documentation in one file. The AI tools beside each article copy the page as Markdown or open it in ChatGPT or Claude. For questions about your own surveys and responses, connect your assistant to Survicate MCP.