Loops
Lightweight, dependency-free, in-memory Loops API fake for testing code that uses the real @loops/loops SDK (or the language-agnostic Loops REST API).
Default port: 4834
Quick start
Start the server:
import { LoopsServer } from "./services/loops/src/server.js";
const server = new LoopsServer(4834);
await server.start();
// ... run your app/tests ...
await server.stop();
Point a client at it (Bearer auth):
await fetch("http://127.0.0.1:4834/v1/transactional", {
method: "POST",
headers: { Authorization: "Bearer parlel-loops-key", "Content-Type": "application/json" },
body: JSON.stringify({
transactionalId: "tmpl_welcome",
email: "recipient@parlel.dev",
dataVariables: { name: "Parlel" },
}),
});
// => { success: true }
Every transactional send is captured and inspectable via /__parlel/*.
Access via MCP / preview URL
When run under the parlel pool, this service is reachable through the MCP gateway
and a preview URL at http://127.0.0.1:4834. Use LOOPS_BASE_URL to point
clients/agents at it. Captured mail lives at GET /__parlel/messages.
Implemented operations
All /v1/* routes require Bearer auth. State is in-memory and ephemeral.
POST /v1/transactional— send a transactional email (transactionalId+emailrequired). Captures and returns{ success: true }.POST /v1/contacts/create— create a contact (emailrequired); returns{ success: true, id }.PUT /v1/contacts/update— update (upsert) a contact by email.GET /v1/contacts/find?email=— find a contact by email (returns an array, empty if none).POST /v1/events/send— send an event (eventName+email/userIdrequired).GET /v1/api-key— validate the API key ({ success: true, teamName }).
Service & inspection operations (parlel extensions)
GET /— service metadata.GET /health— health check ({ status: "ok" }).POST /__parlel/reset— reset all in-memory state.GET /__parlel/messages— list captured transactional emails ({ messages, count }).GET /__parlel/messages/:id— fetch a single captured email.DELETE /__parlel/messages— clear only the captured mailbox.
Surface coverage
This emulator faithfully replicates the API surface most application code and agents exercise. Anything below the supported lines is either an intentional design choice for a fast, zero-cost local emulator (✓ By design) or a candidate for a future release (⟳ Roadmap) — never a silent inaccuracy.
Legend: ✅ fully supported · ◐ accepted (stored, not strictly enforced) · ✓ by design · ⟳ on the roadmap.
| Feature | Status |
|---|---|
sendTransactionalEmail | ✅ Supported |
| Contacts (create/update-upsert/find) | ✅ Supported |
| Events (send) | ✅ Supported |
| API-key validation | ✅ Supported |
| Captured-mail inspection | ✅ Supported (parlel extension) |
| Actual email delivery / SMTP | ✓ By design — Captured in-memory for inspection — no real messages sent |
| Real template rendering / data variables | ◐ Accepted; not rendered |
| Mailing lists / contact properties schema / attachments | ⟳ Roadmap |
| Real API-key validity / scope enforcement | ✓ By design — Any non-empty credential is accepted — no real secrets needed |
Rate limiting (429) | ✓ By design — Never throttles — local tests run at full speed, zero cost |
Error shapes
Errors use the Loops envelope { "success": false, "message": "..." }.
| Status | When |
|---|---|
401 | missing/invalid Bearer auth |
400 | missing transactionalId/eventName, invalid email, malformed body |
409 | contact already exists on create |
404 | unknown endpoint |
Manifest
See services/loops/manifest.json:
- name:
loops, image:parlel/loops:1.0 - port:
4834, protocol:http, healthcheck:/health, startup ≈ 100ms - env:
LOOPS_API_KEY,LOOPS_BASE_URL
Configuration — test.env
Copy these into your test.env (used by the bridge sidecar flow). Tokens are Parlel's seeded test credentials — any non-empty value is accepted by the emulator, so you rarely need to change them. Swap in real credentials only when pointing at the live service in prod.env.
LOOPS_API_KEY=parlel-loops-key
LOOPS_BASE_URL=http://parlel-bridge:4834
<!-- parlel:testenv:end -->