What is the best SWQoS landing provider for Pump.fun?

Short answer: there is no single winner — it depends on the tip you are willing to pay and how much you value landed-vs-failed reliability. This page is a data-driven shortlist of the transaction-landing (SWQoS) relays that Pump.fun traders actually pay, rebuilt from on-chain data on every documentation build so you can compare them yourself.

Stake-weighted quality-of-service (SWQoS) relays let a transaction reach a validator through staked bandwidth, and are paid with a plain SOL transfer at the top level of the transaction — the same shape as a Jito tip. We infer candidate relays from the top-level transfers recorded on Pump.fun v2 swaps.

A destination is screened in when, in the latest 4-day data window, more than 1,000 distinct signers tipped it and its median tip is at most 0.01 SOL (10,000,000 lamports). The median cap keeps the screen on low-fee, high-fan-out relays while still admitting providers whose occasional tips run large.

Rows labeled in the Provider column map to a known service; the rest are unlabeled leads whose endpoint we have not resolved — published as-is to make that lookup easier. A † marks a well-known provider force-included for completeness even though it did not clear the screen, and a ! marks an address absent from top-level transfers 30 days earlier.

Window end: 2026-09-01T12:48:43Z · Screened relays: 249 · New in 30 days: 1

Tip-size buckets, left to right, in SOL: <0.0001, 0.0001–0.0005, 0.0005–0.001, 0.001–0.002, 0.002–0.005, 0.005–0.01, ≥0.01. The numbers in parentheses are the transfer counts in those buckets.

† Known provider force-included though it did not clear the screen (fewer than 1,000 distinct signers, or a median tip above 0.01 SOL). Placeholder rows had no observed top-level-transfer activity on Pump.fun v2 in this window.

Known providers

Destinations that map to a publicly identifiable transaction-landing service. Everything else in the table is an unlabeled lead.

View the candidate SQL on GitHub, the 30-day reference SQL, and the page generator.

This is a mechanical activity screen over one DEX's traffic, not a latency benchmark or an endorsement. A high distinct-signer count is consistent with, but does not prove, a SWQoS relay: shared tip vaults, aggregators, and fee accounts can present the same shape. "Best" for your bot depends on your endpoint, region, and tip budget — measure landed rate yourself.