Configure seine: saved settings and launch options

Make your Blocknet miner start with predictable settings every time. This guide follows seine v0.2.37 and explains where connection settings live, which launch arguments take precedence, and how to make an intentional change persist. Start with an already working installation; the update guide covers replacing the application itself.

How does seine choose its settings?

Find the configuration used by your launcher

The default file is ./seine-data/seine.config.json. That relative path starts at the process working directory, which can differ from the folder containing the executable. Check the terminal folder, shortcut working directory or service settings before editing anything. Two launchers can run the same executable while reading different data folders.

The --data-dir option selects the base directory; the filename remains seine.config.json. Use a stable absolute directory when launching from several places. Record the executable, working directory, data directory and complete arguments together. The Windows startup guide shows how those paths fit into Task Scheduler, including paths containing spaces.

Which settings belong in the saved file?

The example below shows the main connection fields. Replace @yourname with your own registered Blocknet ID or full receiving address. It is a reference for the file shape, not a command to run. When editing an existing installation, change its relevant fields rather than replacing the whole file with this example.

{
  "schema_version": 1,
  "mode": "pool",
  "address": "@yourname",
  "pool_url": "stratum+tcp://bntpool.com:3333",
  "pool_worker": "desktop-01"
}

The fields mode, address, pool_url and pool_worker select mining mode, receiving destination, pool endpoint and rig label. Preserve schema_version and any existing dev_fee_pool_worker field. The latter is a separate generated label; use pool_worker for your own rig. Keep wallet recovery phrases and private keys out of this connection file.

How do you make a change permanent?

Stop the miner before editing and keep a backup of the existing file. Use a plain-text editor, retain valid JSON with double-quoted strings, and save the original filename without an extra text-file extension. Change only the field needed for the task. Save the file in the data directory your regular launcher actually selects. Check commas between fields and avoid comments or trailing commas. If your editor offers JSON validation, use it before restarting. A backup lets you compare the exact previous values instead of reconstructing them from memory.

Inspect that launcher for an explicit override of the same setting. A saved address edit will not take effect while an old --address remains in the command. Either update the launch argument deliberately or remove it so the saved value applies. Restart through the same shortcut, script or task you normally use, then verify the effective settings.

Use temporary overrides deliberately

On an installation with an existing configuration, the following pool rig option selects that rig name for the current run. The saved rig label remains available for the next launch without the override. This is useful for a controlled change, but record the flag in the regular launcher if you want that behavior every time.

--pool-worker desktop-02

On first startup without a configuration file, seine creates one from the initial connection inputs. That bootstrap step differs from overriding an existing file. Keep this distinction in mind when copying a command to another computer: the new machine may be creating its first configuration. The configuration reference documents the first-run prompts and supported launch options.

Keep tuning options with the startup command

Options such as --backend and --threads are supplied as launch arguments. The saved connection schema is not a general list of every flag. Adding a JSON field with a similar name does not set that runtime option. Keep deliberate hardware choices in the script, shortcut or service that starts the miner.

The thread controls in the dashboard change the running session; they do not write a new saved thread default. Automatic tuning also has its own cache files. Preserve those separately from connection edits and let each machine tune for its own hardware. The automatic tuning guide explains cache reuse and the effect of explicit thread choices.

Why do settings seem to reset?

First compare the data path and launch arguments with the last working run. A new extraction directory, another terminal folder or an old scheduled-task argument can explain different settings. If parsing fails, read the filename and message before changing anything else. Restore the backup or correct the JSON syntax, then launch once through the intended entry point.

Confirm the displayed destination, pool and rig name, then look for accepted shares. Repeat a normal restart to check persistence; a single override run only proves that command worked. Use the dashboard guide to read the session and the address-change guide when changing where future pool work is credited. Keep the working launcher and configuration backup together.