Amatya Tests documentation
Everything here is agent-friendly: copy any page as markdown, open it in your LLM, or point an MCP client at the whole platform. Humans welcome too.
Quickstart
From zero to a live test link in ten minutes — by hand or with your AI agent.
Connect your AI agent (MCP)
Point Claude Code, Cursor, or any MCP client at your question bank in two minutes.
JSON import
Paste question JSON from ChatGPT, Claude, or your own scripts — the exact schema.
Marking schemes
Standard and JEE-style partial marking, fully parameterized per section or per question.
Languages
One question, many languages — and tests that refuse to publish with coverage gaps.
Anti-cheat & proctoring
Three levels per test — and an honest account of what each can and cannot do.
API & delegate tokens
Full-access tokens for your own agents; narrow, expiring tokens to let anyone fill your bank.
REST API
The same API the app runs on — cookie sessions for browsers, bearer tokens for machines.
Interactive games
Game-style questions in a locked sandbox — with prompts your favorite LLM can build from.
Embeds & webhooks
Run tests inside your own website and push results into your ERP — the integration rail.
Pricing
Free to start, one-time packs when you need a bigger room, Pro when testing is your rhythm.
Agent-readable everything: /docs/llms.txt