Can your agent
complete a purchase?
AgentData is a testing environment for agentic commerce. Run your agent through realistic buying scenarios — search, compare, checkout — and see whether it actually works.
From zero to testing in 3 steps
No signup required to start. Hit the API directly and build your agent logic against a realistic e-commerce backend.
Connect
Point your agent at the AgentData API. No auth, no setup — just start making requests. Grab the agent guide from /agents.txt to orient your LLM instantly.
Test
Let your agent search products with /api/match, filter by price, accuracy and compatibility, add items to cart, and complete a full checkout flow — just like a real store.
Deploy
Once your agent behaves correctly against AgentData, swap in your production API. Your agent is already battle-tested against real e-commerce flows.
Built for agents, not just humans
The /api/match endpoint is your agent's best friend. Filter the entire catalog with a single query.
Filter by anything that matters
Your agent can narrow down 14 AI products by category, price, latency, accuracy, output format, data region, and platform compatibility — all in one request.
category vision, nlp, code, analytics…max_price max price in USDmax_latency response time in msmin_accuracy score from 0.0 to 1.0region us, eu, globalformat json, text, markdowncompatible_with zapier, n8n, langchain…{ "success": true, "count": 3, "data": [ { "id": "fastsummarizer-pro", "name": "FastSummarizer Pro", "price": 19.99, "accuracy": 0.94, "region": "eu" } ] }
# Find EU NLP agents under $30 with high accuracy curl "https://agentdata.shop/api/match?\ category=nlp&\ region=eu&\ max_price=30&\ min_accuracy=0.9&\ format=json"
# 1. Find a product curl "https://agentdata.shop/api/match?category=code&max_price=60" # 2. Add to cart curl -X POST https://agentdata.shop/api/cart/add \ -H "Content-Type: application/json" \ -d '{"product_id":"codereview-bot","quantity":1}' # 3. Checkout curl -X POST https://agentdata.shop/api/checkout \ -H "Content-Type: application/json" \ -d '{"email":"agent@test.com"}'
# Score your agent against a scenario curl -X POST https://api.agentdata.shop/api/evaluate \ -H "Content-Type: application/json" \ -d '{ "scenario_id": "budget-search", "agent_run": { "product_selected": "destatis-cpi-categories", "price_paid": 7.99, "steps_taken": ["match", "cart/add", "checkout"], "email_provided": "test@example.com", "error_reported": null } }'
Test Scenarios
Run your agent against these scenarios and evaluate whether it actually works.
What agents can test
Real and simulated products for realistic commerce scenarios.
Products marked Real Data deliver real data on purchase. Demo products simulate realistic API responses for testing.
Try it with ChatGPT
Use our AgentData Shopping Assistant to find and purchase products directly from ChatGPT.
Open in ChatGPT →Simple, developer-friendly pricing
Start for free. Scale when you need to. No credit card required to get started.
- ✓ All 5 test scenarios
- ✓ 20 evaluate calls per month
- ✓ Public run history
- ✓
/agents.txtaccess - ✓ No credit card required
- ✓ Unlimited evaluate calls
- ✓ Private run history
- ✓ API Key access
- ✓ Priority support
Angaben gemäß § 5 TMG
Kaeppelstr. 13
90491 Nürnberg