Replay

Replay lets you re-execute previously captured requests — either one at a time or as a saved sequence. This is invaluable for reproducing issues, testing the same request against different servers, or automating repetitive API testing.

How Replay Works

When you replay a request, Madhyamas sends it to the server again using the original method, URL, headers, and body. The response is captured as a new traffic entry, so you can compare it with the original.

Replaying a Single Request

From the Traffic View

1. Right-click any traffic entry in the traffic list
2. Select Replay from the context menu
3. The request is sent immediately
4. The new response appears as a new traffic entry

From the Replay View

1. Navigate to the Replay view
2. Select a previously saved request from the list
3. Click Replay to execute it
4. View the response in the detail panel

Saving Requests for Replay

Instead of searching through traffic history, you can save specific requests for easy replay:

1. Right-click a traffic entry → Save for Replay
2. Give it a name and optional description
3. It appears in the Replay view's saved list

Saved requests persist across restarts, so you can build a library of commonly tested API calls.

Modifying Before Replay (Edit-then-Repeat)

Before replaying, you can modify a saved request and replay it with your changes. The editor is diff-based — it compares your edited values against the original and sends only the fields that changed, so the original request is preserved and only the edited parts are overridden.

From the Web UI

1. In the Saved Requests list, click Edit & Replay next to a saved request
2. The Request Editor dialog opens, pre-filled with the saved request's method, URL, headers, and body
3. Modify any fields:
   • Method — dropdown selector (GET/POST/PUT/PATCH/DELETE/HEAD/OPTIONS)
   • URL — text input
   • Headers — key-value editor with add/remove rows
   • Content-Type — selector that auto-detects from headers
   • Body — monospace textarea
4. Click Replay with Changes. The editor sends only the changed fields as modifications.
5. The replay result dialog shows the response status, headers, and body.

The existing Replay button remains for quick no-edit replay.

From the CLI

The madhyamas replay run command accepts modification flags:

# Simple replay (no modifications)
madhyamas replay run <saved-request-id>

# Override URL and method
madhyamas replay run <id> --url https://staging.example.com/api --method POST

# Add/replace headers (repeatable)
madhyamas replay run <id> --header "Authorization: Bearer token" --header "X-Custom: value"

# Override body from text or a file
madhyamas replay run <id> --body '{"key":"value"}'
madhyamas replay run <id> --body-file ./payload.json

# Follow redirects
madhyamas replay run <id> --follow-redirects

Flag	Description
--url <URL>	Override the request URL
--method <METHOD>	Override the HTTP method
--header "Key: Value"	Add/replace a header (repeatable)
--body <TEXT>	New request body (raw text)
--body-file <PATH>	Read request body from a file
--follow-redirects	Follow 3xx redirect responses

This is useful for:

• Testing different parameter values
• Adding or removing headers
• Changing the request body
• Pointing to a different server

Repeat Advanced (Batch Replay)

Repeat Advanced replays a saved request multiple times with configurable iterations, concurrency, and an optional inter-request delay, then returns aggregate statistics (success/failure counts and latency percentiles). This is useful for basic load testing, performance benchmarking, and verifying endpoint stability under concurrent load.

From the Web UI

1. Open the Replay panel
2. Find a saved request and click the Advanced button
3. Adjust the Iterations slider (1–1000 in the UI)
4. Adjust the Concurrency slider (1–100)
5. Optionally toggle Delay between requests and enter a delay in milliseconds
6. Click Run Batch. A results summary appears with success/failure counts and a latency statistics table (min/avg/max/p95).

From the CLI

# Replay a saved request 100 times, 10 concurrent, 50ms delay
madhyamas replay run-advanced <id> \
  --iterations 100 \
  --concurrency 10 \
  --delay-ms 50

# With modifications (same flags as replay run)
madhyamas replay run-advanced <id> \
  --iterations 50 --concurrency 5 \
  --url https://staging.example.com/api/users \
  --header "Authorization: Bearer token"

Flag	Description	Default
--iterations	Total number of requests to send (max 10,000)	1
--concurrency	Simultaneous in-flight requests (max 100)	1
--delay-ms	Delay between requests in milliseconds	(none)

Result Statistics

The batch result reports:

Field	Description
Total	Total number of requests sent
Succeeded	Requests that completed without error
Failed	Requests that errored
Min / Avg / Max	Latency of successful requests
P95	95th percentile latency of successful requests

TIP

To prevent accidental denial-of-service against upstream servers, iterations are capped at 10,000 and concurrency at 100.

Replay History

Every replay execution is recorded in the Replay History tab. Each entry shows:

• The original request that was replayed
• The timestamp of the replay
• The response status code and timing
• The full response details

This lets you compare results across multiple replays — for example, to see if a server's response has changed over time.

Common Use Cases

Reproducing a Bug

Capture the exact request that caused a bug, save it, and replay it after each code change to verify the fix works.

API Regression Testing

Save a set of key API requests and replay them after deploying a new version to verify the responses haven't changed.

Performance Comparison

Replay the same request multiple times and compare response times to track performance trends.

Testing Different Environments

Save a request, then modify the URL to point to staging or production, and replay to compare responses across environments.

See also

• Breakpoints — modify requests interactively
• Mocks — return fake responses without replaying
• Throttling — replay under simulated network conditions
• CLI reference — madhyamas replay subcommands
• REST API reference — /api/replay endpoints