Simulated environment. Every failure on this page is intentional. seed=42 · deterministic · reqres.in/agent/v1

Built on 12 years · 150M requests/month

The staging internet for agents

Real APIs are dangerous, rate-limited, and never the same twice. ReqRes is the deterministic world where your agent practices: seeded data, deliberate failures, byte-identical replays. The safe place to break things before production.

curl-able in 10 seconds 15 scripted failure modes known to every frontier model

determinism checkseed=42
same seed → same world run it yourself: curl "https://reqres.in/agent/v1/users?limit=1"

Agents need a safe place to fail

Why this exists

Production bites back

An agent practicing retries against a real API is an incident waiting for a timestamp. Rate limits, real money, real customer data. None of it belongs in a training loop.

Mocks rot

Hand-rolled fixtures drift from reality the day you write them. ReqRes serves living, relational, versioned data (users, orders, auth), maintained so you never have to.

Flaky tests teach nothing

If the same request can return different bytes, your eval can't tell a model regression from network weather. Determinism makes agent failures reproducible, and therefore fixable.

A small world, fully specified

GET /agent/v1

/users

Cursor pagination, ULIDs, 20+ fields nested three levels deep, nullable on purpose. 247 users per seed, infinite seeds.

/orders

Relational data that references users, products, and addresses, including line items pointing at deleted products, because real data does that.

/auth/login

Returns a session or an MFA challenge depending on the email. Shape-varying responses, with hypermedia hints for the next step.

/scenarios

Fifteen deliberate failure modes on demand: the catalogue below. Every one tagged X-Agent-Sandbox-Intentional so your error reporter stays quiet.

Trajectory logs

A replayable record of every call your agent made: exportable, diffable, attachable to an eval run. For eval teams, the log is the product.

Agent-native discovery

Point any model at /llms.txt, /llm.txt, or /openapi.json, or hand it the MCP server. No human docs required.

The failure catalogue

15 ways to break, on purpose
free429rate-limitedThrottled, with Retry-After. Does your agent actually wait?
free422validation-errorMulti-field failure with a field_errors object to parse.
free200malformed-jsonA 200 with a truncated body. The cruellest one.
paid500server-errorGeneric 5xx. Exercises the retry path.
paid502bad-gatewayUpstream returned garbage. Test fallback and backoff.
paid504timeoutGateway timeout, simulated delay header included.
paid401unauthorizedBad token, WWW-Authenticate attached.
paid403forbiddenAuthenticated, but missing the scope.
paid404not-foundGone, with a "did you mean?" hint to follow.
paid409conflictOptimistic-concurrency clash: current_version vs yours.
paid413too-largePayload over the limit; max_bytes tells you by how much.
paid200slowSuccess, eventually. Simulated 10s delay.
paid206partial-contentHalf an answer with Content-Range. Handle it.
paid302redirect-loopA 3-hop chain. Exercises follow-redirect caps.
paid200empty-responseContent-Length: 0. Nothing at all, successfully.

Chain them: scripted scenario sequences: succeed twice, then 429, then timeout. Retry logic that survives contact with reality.

Next: simulated services

The roadmap

In build · Q4 2026

Payments sandbox

A Stripe-shaped API where agents practice the scariest integration in software, without touching money.

  • Idempotency keys, enforced
  • Webhook delivery your agent must receive
  • Declined cards, disputes, test clocks
  • Deterministic, seeded, replayable

Queued · 2027

Messaging sandbox

Slack-shaped events, rate limits, and retry semantics for agents that live in channels.

Queued · 2027

CRM sandbox

Salesforce-shaped objects, auth scopes, and query semantics for enterprise agent workflows.

The Agent API Benchmark

Quarterly · public

Which agents handle the real internet?

Every quarter we score the major coding agents against the full failure catalogue (pagination, retries, malformed JSON, auth flows) using trajectory logs from the sandbox. Published in full, methodology open.

AgentPaginationRetry & backoffError handlingOverall
Claude Code
Cursor
Copilot
Devin

First results: September 2026. Get notified when the leaderboard goes live →

Pricing

Start free, no signup

Free

$0

Kick the tyres. No card, no signup.

  • All fixture endpoints
  • 100 requests/day per IP
  • 3 sample failure scenarios
Copy a curl command

Agent Developer

$49/mo

For teams shipping agents that hit APIs.

  • 10M requests/month
  • All 15 failure scenarios + chains
  • Trajectory logs & usage dashboard
  • Issued x-agent-id key
Get Agent Developer

Agent Platform

Custom

For coding-tool companies, eval harnesses, and labs.

  • 100M+ requests/month
  • 90-day trajectory retention
  • Private namespaces & self-hosted image (roadmap, design partners first)
  • SLA & single invoice
Talk to us

Ten design partners. Then we close the doors.

We're onboarding a small group of environment builders, eval teams, and agent companies before public launch: free Agent Platform access through Q1 2027, a direct line on the roadmap, and your logo on this page when we go loud.

5 of 10 slots remaining