Agent-native API
Give your AI agent a survey platform it can actually use
Surveti exposes a hosted Model Context Protocol endpoint, so an AI agent can go from a research question to a live survey and back with analyzed results — the whole loop, over one authenticated connection. Point Claude, Cursor, or any MCP client at it with a scoped API key and it just works.
How it works in Surveti
Speaks MCP, not a bespoke SDK
A single hosted endpoint over the Model Context Protocol. Any MCP-capable client — Claude, Cursor, your own agent — discovers the tools and calls them directly. No client library to install, no schema to hand-write.
The full field-to-result loop
Six focused tools cover the real workflow: list projects and surveys, read a survey, create a draft with its questions, launch it and get a share link, then read analyzed, aggregated results — choice counts, means, NPS, and more.
Scoped keys you control
Create read-only or read-and-write keys per agent, see when each was last used, and revoke any of them instantly. A key carries only the permissions of the workspace it belongs to.
Your governance still applies
Agents work inside the same guardrails as people. Your workspace AI off-switch, plan limits, and audit trail all hold — an API key never routes around them.
What teams use it for
- Have an agent draft and launch a survey from a plain-language brief
- Wire Surveti into an internal research copilot or Slack bot
- Pull analyzed results into an automated readout without exporting by hand
- Let a coding assistant scaffold and field a study while you review
Frequently asked questions
What is the Model Context Protocol?
MCP is an open standard for connecting AI assistants to external tools and data. Surveti runs a hosted MCP server, so any MCP-compatible client can use Surveti as a set of tools without custom integration code.
Which clients can connect?
Any MCP client — including Claude and Cursor — plus your own agents. You add Surveti’s endpoint and an API key to the client’s config, and its tools become available in that session.
What can an agent actually do?
List your projects and surveys, read a survey and its questions, create a draft survey with questions, launch a draft and get its share link, and read analyzed results. Write actions require a key with the write scope.
Is it safe to give an agent a key?
You choose the scope — read-only or read-and-write — and can revoke any key the moment you want. Keys are shown once and stored only as a hash, and every action stays inside your workspace’s permissions, AI off-switch, and audit log.