Questions, and short answers
The data
How fresh is the data?
The archive is 2 to 4 days behind the chain. The order page shows the newest epoch in it. For live data, use a stream from an RPC provider.
What is an epoch?
432,000 slots. In Oct 2026 an epoch lasts about 1.1 days and holds about 0.7 TB of archive. In 2025 it lasted about 2 days.
Which epochs can I scan?
Any epoch from genesis to the newest one in the archive.
Are failed transactions and votes included?
Failed transactions are included by default. Set include_failed to false to drop them. Vote transactions are excluded by default.
Is the data complete?
The scan reads every block in the Old Faithful archive for your range and tests every transaction against your filter. Inner instructions are included. One instruction is one row, and one event is one row.
My program is not an Anchor program.
You still get every matched instruction with its raw bytes and all context columns. Decode the data column in DuckDB, Python or Rust.
My program is not in the presets.
Send its program id in programs, and its Anchor IDL in idl. The 0.29 and 0.30 layouts work, up to 512 KB. See the docs.
How large is the result?
It depends on the filter. One epoch of SPL Token gives about 6.9 GB of Parquet. A narrow account filter can give a few MB. One job delivers 60 GB at most.
The job
How long does a job take?
About 1 to 2 hours for each TB of archive. One recent epoch is about 0.7 TB. The quote shows an estimate, and real scans are often faster.
What if the job fails?
You get a credit for the full price. Send the same job again with X-CREDIT: <credit id>, and you pay nothing. Files that the job uploaded before it stopped stay available for download.
Can a credit become a refund?
Yes. Send the credit id to contact@trawl.run, and we refund it in USDC by hand.
What if I need more than one job can hold?
One job scans 13 TB at most. Send one job for each part of the range. Each job gets its own machine.
Can I scan only part of an epoch?
Yes. Set start_slot and end_slot in the filter. The price is still for whole epochs.
How long do you keep the result?
7 days. Then R2 deletes it. Download it, or copy it to your own storage, in that time.
Payment and trust
How do I pay?
USDC on Solana, through x402. Any x402 client works. Cards are not available now.
Do I need an account or an API key?
No. The payment is the authorization. Keep your job id: it is the key to your result.
Does the filter change the price?
No. The scan reads every byte of the range for any filter. A narrow filter gives a smaller result, not a lower price.
Who can see my filter and my result?
Your filter runs on a machine that exists only for your job. The result is in R2 under your job id. Anybody with the job id or a download URL can read the result, so keep them private.
Are you part of Anza, Triton or the Solana Foundation?
No. This service is independent. It uses the public Old Faithful archive and an open-source fork of Jetstreamer.
How do I contact you?
Write to contact@trawl.run. Add the job id if your question is about a job.
Still have a question?
Write to contact@trawl.run, or get a quote. A quote costs nothing, and it shows the size and the run time of your range.