Do I need OpenSearch running to replay?

No. The recorded responses are served from the test set, so a replay run needs no OpenSearch instance at all. You need it once, during the recording pass.

Is my OpenSearch client still exercised?

Yes. Keploy replays at the protocol, so the client parses, converts and decodes exactly as it does in production. That is the difference from a stubbed client, which skips all of it.

How do I refresh the recordings?

Record again. One command re-captures every OpenSearch interaction, so a schema or response change surfaces as a diff instead of a stale fixture that still passes.

Is scoring stable across replays?

Yes, because the score is part of the recorded response rather than recomputed. That is more stable than a live cluster, where a small mapping change moves scores.

Does Keploy replace Testcontainers?

Not where you need the real engine, and that is worth being clear about. Use Testcontainers when a test genuinely has to exercise OpenSearch itself. Keploy covers the much larger set of tests that only need the responses.

Keploy as a OpenSearch testing framework

Keploy records the HTTP and JSON API exchanges your service has with OpenSearch, then replays them byte for byte in tests. Your OpenSearch client runs unchanged and does its real work on every run, but CI needs no OpenSearch instance, no schema and no seed data.

Generate OpenSearch tests free

keploy record -c "<your app start command>"

18.5K+

1.2M+300M+mocks createdSlack

What Keploy gives a OpenSearch team

Every OpenSearch call your service makes

Keploy records OpenSearch search and index responses over HTTP, so a service that queries an index can be tested with no cluster, no mapping and no refresh interval to wait on.

Search and index responses captured
Aggregations captured
Real scoring in the response
No cluster in CI

Try Keploy for free Explore the docs

The problem

Why OpenSearch integration tests slow teams down

The friction is rarely the assertions. It is everything around them — spinning up dependencies, keeping mocks honest, and repairing tests after every refactor.

Book a call Explore the docs

Three ways to do it

Testing OpenSearch: by hand, with Testcontainers, or with Keploy

Testcontainers gives a test a real OpenSearch node at the cost of running one. A hand-written stub is cheap but never exercises the search API. Keploy replays the recorded exchange with nothing running.

By hand

Testcontainers

With Keploy

Select any row for the full comparison, with code.

Get started Book a call

Same coverage, three costs

What you write for OpenSearch vs what Keploy records

All three produce the same assertion. Only the third still passes after the next refactor without anyone editing it.

By hand2–4 hours

order_repository_test.py

hand-written

from unittest.mock import MagicMock

# A stub OpenSearch client: every reply is typed out here.

def test_reads_an_order():

client = MagicMock()

client.fetch.return_value = {

"id": "o-1",

"total": 4200,

"status": "NEW",

}

order = OrderRepository(client).find("o-1")

assert order.total == 4200

# Nothing here exercised the driver, so a type or encoding

# change in the real reply cannot fail this test.

A hand-written stub of the OpenSearch client. The driver never runs.

Testcontainers1–2 hours

order_repository_container_test.py

tool-assisted

import pytest

from testcontainers.core.container import DockerContainer

@pytest.fixture(scope="session")

def client():

# Boots a real OpenSearch per suite. Needs a Docker daemon

# available wherever the tests run, including CI.

with DockerContainer("opensearch:latest") as c:

c.start()

yield connect(c.get_container_host_ip(), c.get_exposed_port(0))

def test_reads_an_order(client):

# Schema and seed data are still written by hand.

seed(client, id="o-1", total=4200, status="NEW")

order = OrderRepository(client).find("o-1")

assert order.total == 4200

Testcontainers runs the real OpenSearch. Fidelity is genuine; the job pays for it.

With Keploy~5 minutes

keploy/test-set-0/test-1.yaml

auto-generated

version: api.keploy.io/v1beta1

kind: Http

name: test-1

spec:

req:

method: POST

url: /orders/o-1/confirm

resp:

status_code: 200

body: '{"id":"o-1","status":"CONFIRMED","total":4200}'

assertions:

noise:

- body.confirmedAt

---

kind: Http

spec:

request:

url: https://api.payments.example.com/v1/charges

response:

body: '{"status":"AUTHORIZED"}'

---

kind: Http

spec:

request:

url: https://api.payments.example.com/v1/payment_intents/tok_123/confirm

response:

body: '{"status":"AUTHORIZED"}'

Recorded from one real request. Nothing here was typed by hand.

Times are estimates for authoring one endpoint’s coverage from scratch, not measurements.

Explore the docs Request a trial

Keploy vs the alternatives

OpenSearch testing tools, compared

The options a team on caches and search stores actually reaches for, and where each one genuinely wins. Select a row for the full comparison.

Keploy

Testcontainers

fakeredis

miniredis

redis-mock

Postman

SmartBear ReadyAPI

Best in classStrongPartialNot covered

Assessments reflect each tool’s documented behaviour, not benchmark measurements.

Request a trial Book a call

How it works

Record your OpenSearch app once, replay it forever

Keploy sits below your OpenSearch process at the network layer. It watches the calls your app already makes, then serves them back on replay so tests run with no dependencies attached.

Your OpenSearch appGET/api/v1/orders/{id}200

RecordingKeploy proxyeBPF · userspace

JavaOpensearch-java

PythonOpensearch-py

Node.js@opensearch-project/opensearch

GoOpensearch-go

ElasticsearchThe same HTTP request shape

An example shape of a captured call. Your own endpoints and dependencies come from your real traffic, so nothing here has to be written by hand.

Try Keploy for free Request a trial

Mock coverage

What Keploy mocks for OpenSearch, with zero config

5 of the 5 dependencies a typical OpenSearch service talks to are stubbed from the recording itself — no mock classes, no fixture files, no containers in CI.

Explore the docs Book a call

Quick start

Your first OpenSearch test suite in under five minutes

Every command below runs against your existing OpenSearch service. Nothing in your source tree changes.

1
1Install the Keploy CLI
A single binary. It needs a Linux kernel with eBPF support, or Docker on macOS and Windows — and it adds nothing to your project's dependencies.
curl -sSL https://keploy.io/install.sh | bash
2
2Record your service
Pass the command you already use to start the app. Keploy runs it and watches every socket it opens.
keploy record -c "<your app start command>"
3
3Exercise the paths you care about
Use curl, your frontend, or an existing smoke script. Every request becomes a test case with its OpenSearch calls captured alongside it.
curl -X POST localhost:8080/orders/o-1/confirm
curl localhost:8080/orders/o-1
4
4Replay in CI
Replay serves the recorded dependency responses, so the job needs no service containers and no Docker daemon.
keploy test -c "<your app start command>" --delay 10