# Launch your first sniper bot

_Source: https://docs.neurotrading.io/zh/tutorials/launch-a-sniper_

Goal: a conservative Pump.fun launch sniper with strict filters and automatic exits.

## Prerequisites

* A **dedicated wallet** for sniping (create one — don't snipe from your main wallet) funded with SOL for several buys plus fees.

## Step 1 — Create the bot

On the web, open **Sniper bot** → **New**. Name it, pick the sniper wallet, and choose the profile — for this tutorial, **Pump.fun launches**.

## Step 2 — Size the buys

* **Buy amount:** small — e.g. 0.05 SOL. You will lose some races and some tokens will rug; size so ten losses don't hurt.
* **Gas / Jito tip:** above default — snipes are races, and a buy that lands late often buys the top. The Jito bundle route also gives MEV protection.

## Step 3 — Set strict filters

For a first bot, require a lot:

* ✅ Twitter/X **and** Telegram present
* ✅ CA ↔ Twitter handle match (dodges impersonators)
* ✅ Unique ticker/name (skips copycats)
* Optional: KOL involvement above a SOL threshold — fewer, higher-conviction snipes

Strict filters mean the bot fires rarely. That's the point — loosen gradually once you trust it.

## Step 4 — Attach exits

Give every sniped position an automatic exit:

* **Take-profit:** e.g. +100%, sell 100%.
* **Stop-loss:** e.g. −40%.
* **Timed auto-sell** as a backstop (e.g. a few hours), so nothing is held forever by accident.

## Step 5 — Arm it and watch

Enable the bot. Each snipe arrives as a Telegram notification with one-tap actions (**Sell 50% / Sell initial / Token info**). The bot's page shows its stats — attempts, wins/losses, realized PnL — and current holdings.

## Variations

* **Graduation sniping:** same setup with the graduation profile — buys tokens completing their bonding curve, generally less rug-prone than fresh mints but more competed.
* **Specific address:** know a token pre-graduation? `/sniper` → buy specific address pre-arms a buy that fires the moment it migrates.

> **Warning**
>
> Review results after every session before increasing size. If the bot's TP/SL settings change later, fires use the **current** config — edits apply to already-open positions' exits too.