Flashpoint.AIFlashpoint.AIdocs

Introduction

Flashpoint.AI is a research platform you operate through chat. You describe a research objective in natural language. The platform plans the work, recruits respondents, fields the survey, analyzes the data, and returns the artifacts.

Two principles shape everything in these docs: the platform is open, and the research it produces is rigorous.

Open by design

No part of the platform is walled off:

  • Open to humans and agents. The chat UI, the MCP server, and the REST API expose the same capabilities. An autonomous agent can run a complete study end to end.
  • Open to your tools. Plug third-party tools in over MCP and the API — or pull Flashpoint.AI capabilities into your own product.
  • Open on the way out. Export surveys to Qualtrics, Confirmit, and Forsta; export data to CSV and XLSX. Your work stays portable.
  • Open pricing. Usage-based and transparent. Agents pay per call over x402.
  • Open to any audience. Market Experiments measure real behavior from people who will never take a survey — research is not limited to panels.

Rigorous by default

AI makes research fast. The platform's job is to keep it sound. Structure is enforced, not suggested:

  • SurveyCheck walks every path before a survey fields. It proves each quota cell is reachable, surfaces unanswerable questions, and gates publishing on an unacknowledged failing check.
  • Data quality is screened on every response. Research Defender and ReDem validate paid sample. Speeder and straight-liner detection runs on every survey. ReCaptcha blocks bots.
  • Synthetic sample is measured, not assumed. Every synthetic panel carries a Fit Score against its target population, and calibrations are validated with k-fold testing.
  • Analysis is statistical. Crosstabs carry significance tests. Surveys are versioned. Every failure mode has a documented error code.

If your agent runs research here, it inherits this structure. The same checks gate an agent's study and a human's — speed without slop.

Three ways in

SurfaceWho it's forWhere
Chat UIResearchers, PMs, marketers — humans doing studiesapp.flashpoint.ai
MCP serverAutonomous agents that pay per call (USDC over x402) and humans driving Flashpoint.AI from Claude Desktophttps://api.flashpoint.ai/mcp
REST APIServer-side integrations and back-office automation against the surveys serviceSee API Reference

The chat UI, the MCP server, and the REST API share the same domain model and the same data. You can start a study in chat and analyze it through the API, or have an agent run a panel and review the results in the UI.

What it covers

  • Surveys — design with 20+ question types, skip logic, quotas, templates; lifecycle from draft to complete with versioning. Read more.
  • Sample — AI-generated synthetic respondents, or real panels from Prolific and Dynata. Each source tagged for per-panel analysis. Read more.
  • Distribute — public links, tracked email lists, panel-provider recruitment. Read more.
  • Analyze — frequencies, crosstabs, NPS, chi-square, segment filters, AI-powered insights. Read more.
  • Deliver — DOCX import, CSV / XLSX / Qualtrics / Confirmit / Forsta exports, presentation generation. Read more.

Design principles

  1. Chat is the primary surface. Every action available in the API is also available through chat. The agent plans multi-step workflows, reports progress as it works, and requests approval before destructive actions.
  2. All channels share one model. A survey written through chat, the MCP tool, or the REST API reads back identically through any of them.
  3. Real and synthetic respondents share a schema. Responses are tagged by source, so you can analyze them together or separately.
  4. Errors are predictable. Every failure mode has a documented code and an HTTP status. See Errors.

For agents

Every page in these docs is available as raw markdown. Replace the path prefix with /raw to fetch the source:

GET https://docs.flashpoint.ai/quickstart       → rendered HTML
GET https://docs.flashpoint.ai/raw/quickstart   → raw markdown (text/markdown)

Or click View as markdown at the top of any page.

What's here

  • Quickstart — first tool call in under five minutes
  • Connect from your AI assistant — per-client setup for Claude Desktop, Claude.ai web, Claude Code, and ChatGPT
  • Authentication — three auth flavors: chat UI (Auth0), MCP for autonomous agents (wallet JWT), MCP for humans (OAuth via Claude Desktop)
  • API Reference — every MCP tool, every request and response
  • MCP capability reference — all ~150 things the surveys agent can do over MCP, what it cannot, and the limits
  • Errors — error codes and how to recover

Conventions

  • Code samples default to curl. Tabs for language alternatives are coming.
  • Field names in prose are wrapped in backticks: team_id, Authorization.
  • Required parameters are flagged inline. Optional parameters note their default.