> For the complete documentation index, see [llms.txt](https://docs.agentipy.fun/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.agentipy.fun/tools/use-pump.fun.md).

# Use Pump.fun

The `Use Pump.fun` tool, part of AgentiPy, allows users to interact with the Pump.fun platform on Solana. This includes functionalities to buy and sell tokens listed on Pump.fun, and to get information about a token's bonding curve state and price. Pump.fun is known for facilitating the launch and trading of new tokens, often with high volatility.

## Functionality

This tool is implemented in the `PumpfunManager` class. It interacts directly with the Pump.fun on-chain program (`PUMP_PROGRAM`) by constructing and sending Solana transactions.

### Key Methods:

* `get_pump_curve_state(conn: AsyncClient, curve_address: Pubkey) -> BondingCurveState`:
  * (Static method) Fetches and parses the account data for a given `curve_address` (bonding curve) to determine its current state (e.g., virtual token reserves, virtual SOL reserves).
  * Returns a `BondingCurveState` object.
* `calculate_pump_curve_price(curve_state: BondingCurveState) -> float`:
  * (Static method) Calculates the current price of the token based on the `curve_state`.
  * Returns the price in SOL per token.
* `get_token_balance(conn: AsyncClient, associated_token_account: Pubkey) -> int`:
  * (Static method) Fetches the balance of a given `associated_token_account`.
  * Returns the balance as an integer (in the smallest unit of the token).
* `buy_token(agent: SolanaAgentKit, mint: Pubkey, bonding_curve: Pubkey, associated_bonding_curve: Pubkey, amount: float, slippage: float = 0.01, max_retries=5)`:
  * (Static async method) Buys a token from its Pump.fun bonding curve.
  * Parameters:
    * `agent`: The `SolanaAgentKit` instance.
    * `mint`: The `Pubkey` of the token to buy.
    * `bonding_curve`: The `Pubkey` of the token's bonding curve account.
    * `associated_bonding_curve`: The `Pubkey` of the curve's associated token account.
    * `amount`: The amount of SOL to spend.
    * `slippage`: Allowed slippage (e.g., 0.01 for 1%).
    * `max_retries`: Number of times to retry the transaction if it fails.
  * Calculates the expected token amount and maximum SOL to spend with slippage.
  * Creates the buyer's Associated Token Account (ATA) if it doesn't exist.
  * Constructs and sends a transaction to the Pump.fun program to execute the buy.
  * Returns the transaction signature.
* `sell_token(agent: SolanaAgentKit, mint: Pubkey, bonding_curve: Pubkey, associated_bonding_curve: Pubkey, slippage: float = 0.25, max_retries=5)`:
  * (Static async method) Sells a token back to its Pump.fun bonding curve.
  * Parameters are similar to `buy_token`, but `amount` is not directly passed; instead, it fetches the full token balance from the agent's ATA to sell.
  * Calculates the minimum SOL output to receive based on slippage.
  * Constructs and sends a transaction to the Pump.fun program to execute the sell.
  * Returns the transaction signature.

##

**Important Considerations:**

* **Direct On-Chain Interaction:** This tool interacts directly with the Solana blockchain and the Pump.fun smart contract. It does not use a proxy.
* **Extreme Financial Risk:** Pump.fun is a platform for newly launched tokens which are often extremely volatile and can result in significant or total loss of funds. This is not an endorsement or recommendation to use Pump.fun.
* **Mainnet Operations:** Pump.fun primarily operates on Solana Mainnet. All transactions are real and involve real SOL and tokens.
* **Slippage:** High slippage tolerance is often necessary due to the volatility of tokens on Pump.fun, which further increases risk.
* **Associated Token Accounts:** The tool attempts to create the Associated Token Account (ATA) for the buyer if it doesn't exist.

## Source Code

You can find the source code for this tool on GitHub: <https://github.com/niceberginc/agentipy/blob/main/agentipy/tools/use_pumpfun.py>