diff --git a/package.json b/package.json
index 66a1cc8..2b5a417 100644
--- a/package.json
+++ b/package.json
@@ -11,6 +11,7 @@
     "deploy:contract": "./scripts/deploy.sh",
     "gen:bindings": "./scripts/generate-contract-bindings.sh",
     "changelog:update": "conventional-changelog -p angular -i CHANGELOG.md -s",
+    "load-test": "k6 run scripts/load-test/stream-load-test.js",
     "prepare": "husky"
   },
   "license": "MIT",
diff --git a/scripts/load-test/README.md b/scripts/load-test/README.md
new file mode 100644
index 0000000..8207446
--- /dev/null
+++ b/scripts/load-test/README.md
@@ -0,0 +1,87 @@
+# Load Tests
+
+This directory contains load/performance tests for the StellarStream backend API.
+
+## Prerequisites
+
+- [k6](https://k6.io/docs/getting-started/installation/) installed and on your `PATH`.
+- The backend server running locally (or a deployed instance).
+
+## Running
+
+Start the backend (see `backend/README.md` or the root `README.md` for setup), then:
+
+```bash
+k6 run scripts/load-test/stream-load-test.js
+```
+
+Or via the root package script:
+
+```bash
+npm run load-test
+```
+
+### Configuration
+
+The script reads the following environment variables:
+
+| Variable     | Default                    | Description                                        |
+| ------------ | -------------------------- | -------------------------------------------------- |
+| `BASE_URL`   | `http://localhost:3001`    | Base URL of the backend API.                       |
+| `JWT_SECRET` | `test_secret_for_load_test`| Secret used to sign JWTs. Must match the backend's `JWT_SECRET` so the generated tokens are accepted. |
+| `VUS`        | `200`                      | Maximum number of concurrent virtual users.        |
+| `DURATION`   | `5m`                       | Duration of the sustained-load phase.              |
+
+Example with a custom target:
+
+```bash
+BASE_URL=http://localhost:3001 JWT_SECRET=my-secret VUS=200 DURATION=5m \
+  k6 run scripts/load-test/stream-load-test.js
+```
+
+## Covered Endpoints
+
+- `POST /api/streams` — create a stream
+- `GET /api/streams` — list streams
+- `POST /api/streams/:id/cancel` — cancel a stream
+
+## Load Profile
+
+The test ramps from **10** concurrent users up to the configured maximum
+(default **200**), holds that load for the configured duration, then ramps
+back down to zero.
+
+## SLOs
+
+The following SLOs are enforced via k6 thresholds. The test **fails** if any
+threshold is breached:
+
+| SLO                          | Threshold          |
+| ---------------------------- | ------------------ |
+| p95 latency (all endpoints)  | `< 200 ms`         |
+| Error rate (all endpoints)   | `< 0.5%`           |
+
+## Report
+
+k6 emits a summary report at the end of the run with a full percentile
+breakdown (`p50`, `p90`, `p95`, `p99`, `max`) for each covered endpoint via
+custom `Trend` metrics:
+
+- `http_req_duration_create`
+- `http_req_duration_list`
+- `http_req_duration_cancel`
+
+For a JSON report, add `--summary-export=report.json`:
+
+```bash
+k6 run --summary-export=report.json scripts/load-test/stream-load-test.js
+```
+
+## Data Integrity
+
+The test creates a stream and immediately cancels it in the same iteration,
+so no orphaned streams accumulate. Each virtual user uses a distinct JWT, and
+the test asserts on HTTP status codes to detect unexpected failures. To verify
+no data corruption under load, inspect the backend logs and the stream list
+after the run (e.g. `GET /api/streams`) to confirm all created streams were
+either cancelled or are in a valid state.
diff --git a/scripts/load-test/stream-load-test.js b/scripts/load-test/stream-load-test.js
new file mode 100644
index 0000000..da883ea
--- /dev/null
+++ b/scripts/load-test/stream-load-test.js
@@ -0,0 +1,175 @@
+#!/usr/bin/env k6
+/**
+ * Load test for stream create, list, and cancel endpoints.
+ *
+ * Usage:
+ *   k6 run scripts/load-test/stream-load-test.js
+ *
+ * Environment variables:
+ *   BASE_URL       Base URL of the backend API (default: http://localhost:3001)
+ *   JWT_SECRET     Secret used to sign JWTs (must match the backend's JWT_SECRET)
+ *   VUS            Max concurrent users (default: 200)
+ *   DURATION       Test duration (default: 5m)
+ *
+ * SLOs (enforced via thresholds):
+ *   - p95 latency < 200ms for all covered endpoints
+ *   - error rate < 0.5%
+ *
+ * The script generates a valid JWT locally using the same JWT_SECRET the
+ * backend uses, so no external Horizon interaction is required. It ramps
+ * from 10 to the configured max concurrent users.
+ */
+
+import http from "k6/http";
+import { check, sleep } from "k6";
+import { Trend, Rate, Counter } from "k6/metrics";
+import { SharedArray } from "k6/data";
+import crypto from "k6/crypto";
+
+const BASE_URL = __ENV.BASE_URL || "http://localhost:3001";
+const JWT_SECRET = __ENV.JWT_SECRET || "test_secret_for_load_test";
+const MAX_VUS = Number(__ENV.VUS || 200);
+const DURATION = __ENV.DURATION || "5m";
+
+// ---- Custom metrics for percentile breakdown per endpoint ----
+const createLatency = new Trend("http_req_duration_create", true);
+const listLatency = new Trend("http_req_duration_list", true);
+const cancelLatency = new Trend("http_req_duration_cancel", true);
+const createErrors = new Rate("create_errors");
+const listErrors = new Rate("list_errors");
+const cancelErrors = new Rate("cancel_errors");
+const streamsCreated = new Counter("streams_created");
+const streamsCancelled = new Counter("streams_cancelled");
+
+// ---- SLO thresholds ----
+export const options = {
+  scenarios: {
+    ramp_up: {
+      executor: "ramping-vus",
+      startVUs: 10,
+      stages: [
+        { duration: "1m", target: Math.round(MAX_VUS / 2) },
+        { duration: "1m", target: MAX_VUS },
+        { duration: DURATION, target: MAX_VUS },
+        { duration: "30s", target: 0 },
+      ],
+      gracefulRampDown: "30s",
+    },
+  },
+  thresholds: {
+    // SLO: p95 latency < 200ms for every covered endpoint
+    http_req_duration_create: ["p(95)<200"],
+    http_req_duration_list: ["p(95)<200"],
+    http_req_duration_cancel: ["p(95)<200"],
+    // SLO: error rate < 0.5%
+    create_errors: ["rate<0.005"],
+    list_errors: ["rate<0.005"],
+    cancel_errors: ["rate<0.005"],
+    // Overall error rate must stay under 0.5%
+    http_req_failed: ["rate<0.005"],
+  },
+};
+
+// ---- Deterministic test accounts (valid Stellar public keys) ----
+const SENDER = "GCLJJD5FHTSEBHFXAA3BBTODUJ4RXDK6B3OSVNKY6TUHF76AQQT2WNFC";
+const RECIPIENT = "GDL7R4FQ6XQ7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7Q7";
+
+// ---- JWT generation (mirrors backend/src/services/auth.ts) ----
+// The backend signs JWTs with HS256 using JWT_SECRET. We replicate that here
+// so the load test can mint valid tokens without hitting Horizon.
+function base64url(input) {
+  return Buffer.from(input).toString("base64url");
+}
+
+function signJwt(accountId) {
+  const header = base64url(JSON.stringify({ alg: "HS256", typ: "JWT" }));
+  const now = Math.floor(Date.now() / 1000);
+  const payload = base64url(
+    JSON.stringify({
+      accountId,
+      iat: now,
+      exp: now + 3600, // 1 hour
+    }),
+  );
+  const signingInput = `${header}.${payload}`;
+  const signature = crypto.hmac("sha256", JWT_SECRET, signingInput, "base64url");
+  return `${signingInput}.${signature}`;
+}
+
+// ---- Pre-generate a pool of JWTs (one per virtual user) ----
+const tokens = new SharedArray("tokens", function () {
+  const arr = [];
+  for (let i = 0; i < MAX_VUS; i++) {
+    arr.push(signJwt(SENDER));
+  }
+  return arr;
+});
+
+// ---- Payload helpers ----
+function createStreamPayload() {
+  const now = Math.floor(Date.now() / 1000);
+  return JSON.stringify({
+    sender: SENDER,
+    recipient: RECIPIENT,
+    assetCode: "USDC",
+    totalAmount: 1000 + Math.floor(Math.random() * 9000),
+    durationSeconds: 3600 * 24 * 30, // 30 days
+    startAt: now + 60,
+  });
+}
+
+function authHeaders(token) {
+  return {
+    Authorization: `Bearer ${token}`,
+    "Content-Type": "application/json",
+  };
+}
+
+// ---- Main test ----
+export default function () {
+  const token = tokens[__VU - 1];
+  const headers = authHeaders(token);
+
+  // 1) POST /api/streams — create a stream
+  const createRes = http.post(
+    `${BASE_URL}/api/streams`,
+    createStreamPayload(),
+    { headers },
+  );
+  createLatency.add(createRes.timings.duration);
+  const createOk = check(createRes, {
+    "create returns 201": (r) => r.status === 201,
+  });
+  createErrors.add(!createOk);
+  if (createOk) {
+    streamsCreated.add(1);
+  }
+
+  // 2) GET /api/streams — list streams
+  const listRes = http.get(`${BASE_URL}/api/streams?limit=20`, { headers });
+  listLatency.add(listRes.timings.duration);
+  const listOk = check(listRes, {
+    "list returns 200": (r) => r.status === 200,
+  });
+  listErrors.add(!listOk);
+
+  // 3) POST /api/streams/:id/cancel — cancel the stream we just created
+  if (createOk) {
+    const streamId = createRes.json("data.id");
+    const cancelRes = http.post(
+      `${BASE_URL}/api/streams/${streamId}/cancel`,
+      null,
+      { headers },
+    );
+    cancelLatency.add(cancelRes.timings.duration);
+    const cancelOk = check(cancelRes, {
+      "cancel returns 200": (r) => r.status === 200,
+    });
+    cancelErrors.add(!cancelOk);
+    if (cancelOk) {
+      streamsCancelled.add(1);
+    }
+  }
+
+  sleep(1);
+}