Pump.fun Volume Bot: How to Run One Without Wasting the Launch

A Pump.fun volume bot is software that funds a fleet of disposable Solana wallets and pushes buy and sell orders through them on a schedule, so a token shows continuous trade flow on the Pump.fun board instead of a flat line. It moves the numbers the board sorts on - traded volume, transaction count, distinct makers, reply activity. It does not manufacture demand, and no bot can promise a trending slot.

What a volume bot actually does

Strip the marketing away and the mechanic is simple. The bot derives a set of fresh keypairs, funds each one with a small, uneven amount of SOL, and then sends swap instructions against the token's bonding curve from those wallets, one after another, with the sizes and the gaps randomized. Each swap is an ordinary transaction. Anyone can see it. The point is not to hide the trades - it is to make the aggregate look like a room full of people rather than one person clicking fast.

On Pump.fun the curve itself does the pricing. Buys push the price up along the curve, sells push it back down, and the fee on each swap is taken by the venue. That is the first thing operators get wrong: a bot that buys and sells the same notional back and forth does not lift the price. It lifts the volume counter and pays the venue for the privilege. Net buy pressure is what moves the chart, and net buy pressure costs real inventory.

So what are you actually buying? Time on the board. Sorting rank. A tape that a human scanning fifty new mints will stop on. A reply feed that is not empty. If a token is good, that window is the difference between finding an audience and never being seen. If a token has nothing behind it, the window closes and the chart goes back to where it was, minus your fees. That is not cynicism, it is just how the mechanism works, and any tool that tells you otherwise is selling you a story.

What the Pump.fun board reads

Pump.fun has never published its ranking formula, and it has been changed more than once. What is observable from the front end is which signals it surfaces: how recently a token launched, how much volume it has done, how many transactions and how many separate wallets are behind them, how busy the reply feed is, and roughly where it sits on the curve. Anyone quoting you an exact market-cap threshold or a precise volume target is repeating a number they saw in a Telegram group.

The practical read is relative, not absolute. You are not competing with a threshold, you are competing with whatever else launched in the same hour. On a quiet weekday morning a modest session stands out. During a mania, the same session disappears. This is why sensible operators size a session against the board they are actually looking at, not against a number from a blog post written six months ago.

Two signals get neglected because they are boring. The first is maker count: fifty wallets doing 500 SOL reads very differently from two thousand wallets doing the same 500 SOL. The second is the reply feed. A chart with movement and a dead comment section tells a visitor that nobody is home, and it is the cheapest of all the signals to keep alive.

The real cost of a session

Here is the part most guides skip. A volume session has four cost lines, and only one of them belongs to the tool you are paying.

Put numbers on it. Take the smallest session we run: 100 SOL of routed volume across 500 wallets, orders between 0.10 and 0.30 SOL, spread over an hour. Our fee is 1 SOL, flat. The Solana side is almost nothing - a few thousand signatures at 0.000005 SOL each comes to roughly 0.01 SOL. Rent is the line that surprises people: 500 fresh token accounts at 0.00203928 SOL is about 1.02 SOL parked in the network, and it comes back when those accounts are closed at the end of the session, which is exactly why closing them is part of the job and not an afterthought.

The venue fee is the one that scales with how hard you churn. Every swap pays it, in both directions, so a strategy built on rapid round trips burns a percentage of the notional every lap. This is the honest argument for spreading a session out instead of hammering it: fewer, better-placed trades over a longer window produce a more convincing tape and pay the venue less than the same volume slammed through in fifteen minutes.

Running a session, step by step

1. Paste the mint address

   Open the console and paste the Pump.fun contract address. The name, symbol and image are read back from public on-chain metadata so you can confirm you are funding the right mint before anything is paid.

2. Pick a size

   Choose a preset - 100, 300, 500 or 1000 SOL of routed volume - or set the wallet count and order band by hand. The wallet count decides how many distinct makers appear; the order band decides how the sizes scatter.

3. Set the session length

   Spread the same volume over 15 minutes for a spike or 10 hours for a slow build. Length is the single setting that changes how the tape reads, because it sets the gap between trades.

4. Turn on the engagement layer

   Enable auto-comments and auto-favorites at the density you want, and optionally have buying wallets follow the token creator. Comments are drawn from a multilingual library, or from a .txt file of your own lines that never leaves your browser.

5. Fund the session

   Send the quoted amount - a flat 1% of target volume - to the address shown, from any wallet. The console watches that address on-chain and starts as soon as the transfer confirms.

6. Watch it land and follow the migration

   Track filled orders, makers, comments and favorites while the session runs. When the curve completes, routing follows the token into its AMM pool instead of stopping at the migration.

The whole thing takes about a minute of actual clicking. The thinking happens before: deciding how big the session should be for the board you are launching into, and how long you need to stay visible.

2:03

• 300 SOLtarget volume
• 600rotating wallets
• 960orders
• 3.6 SOLtotal fee

1. 0:00Why a launch stays invisible
2. 0:12Paste the mint and resolve it on chain
3. 0:21Wallets and the order band you choose
4. 0:41Target volume, session length and pacing
5. 0:49The comment library and your own file
6. 1:02Creator follows
7. 1:06Session summary and how the fee is worked out
8. 1:27Funding with Solana Pay
9. 1:32Payment confirmed on chain
10. 1:35The volume session running
11. 1:47Session finished and logged

The run in that recording is deliberately mid-sized. Six hundred wallets at an average of half a SOL each is enough spread to read as a crowd rather than a handful of accounts, and three hours is long enough that the tape survives a timezone change without the session outliving the attention it buys.

Settings that hold up

Wallet count sets your maker count. This is the setting people under-use because it is invisible in the volume number. If you are choosing between more wallets with smaller orders and fewer wallets with larger ones, take the wallets. Distinct makers is a harder signal to fake after the fact, and it is what separates a crowd from a whale.

The order band is your texture. A band of 0.10 to 0.30 SOL gives you a tape where no two prints look alike. A band of 0.20 to 0.21 gives you a barcode. Widen the band when you want occasional larger prints to punctuate the flow; keep it tight and low when you want the session to look like patient accumulation.

Length is the real strategy setting. Fifteen minutes is a spike. An hour is a launch window. Five hours is a build. Ten hours is a slow accumulation that will still be running when the timezone flips and a different audience opens the board. Same volume, four completely different stories.

Comment density is not decoration. Around 40% of orders carrying a reply keeps a feed alive without turning it into spam. Every comment posts from a different wallet, in mixed languages, because a reply feed where every line is in the same register reads worse than no replies at all. If you have your own lines, load a .txt file - in our console that file is parsed in the browser and never uploaded anywhere.

Favorites and creator follows are the cheap signals. They cost almost nothing and they fill in the part of the profile a visitor checks after the chart: does anyone care about this token beyond the trade.

Five ways operators waste a session

1. Funding every wallet from one address in one block. The funding graph is public. A hundred wallets that all received exactly 0.4 SOL from the same source in the same slot is the single most obvious pattern on the chain.
2. Running the session before the token is ready. No image, three lines of description, no social. The volume works, people arrive, the page tells them nobody is home, and the fee is gone.
3. Buying volume instead of buying attention. Chasing a big volume number with fast round trips maximises venue fees and minimises how long you stay visible. Stretch the same budget and you get more of what you actually want.
4. Stopping at the migration. The curve completes, liquidity moves to an AMM pool, the token gets its widest audience of the day, and the flow stops dead. Routing should follow the token across.
5. Promising a price. The fastest way to turn a marketing spend into a legal problem is to tell buyers what the chart will do. Do not.

Volume bot, bundler, sniper, market maker

These four get used interchangeably in Telegram and they are not the same tool.

How volume bots are priced

Three models are common. Per-maker pricing charges a fixed amount for each batch of wallets, which is easy to quote and hard to compare, because the thing you care about - how much volume actually gets routed - is not what you are paying for. Subscription pricing charges monthly whether you launch or not. Percentage pricing charges against the volume you configure.

We use the third, at a flat 1% of target volume with a 100 SOL minimum, because it is the only one of the three where the invoice moves with the work. A 100 SOL session costs 1 SOL. A 1,000 SOL session costs 10 SOL. Nothing is bolted on afterwards for comments, favorites, priority fees or the Jito tips, and the exact figure is on screen before you send anything. Whatever tool you end up using, ask the same question: what is the all-in number, and what is not included in it.

Risk, ethics and what we will not claim

Volume routing sits in an area that regulators have started paying attention to. In October 2024, US prosecutors charged several firms that had been selling market-making services to token projects, and the allegation was wash trading - trading with yourself to create the appearance of a market. The detail that matters for anyone reading this: the charges were about representing that activity as organic demand. Marketing your own flow as other people's buying is where the line sits in most jurisdictions, and it is a line you can cross with a tweet.

So, plainly: this is a visibility tool. We do not claim it produces a trending slot, a price, a holder base or a profit. We do not know your local law and this is not legal advice. What we will state is what the session does on-chain, what it costs, and what settles, because those are the only parts anyone can verify. If a service is willing to promise you the rest, that willingness is the warning.

Questions people actually ask

Kristjan Kask, Founder

Builds and runs the PumpWave session engine at PumpWave Labs in Estonia. Writes about Solana launch mechanics from the operator side: what settles, what it costs, what the board does with it. Corrections and arguments to support@pumpwave.net.