The limits are the sources' limits
This software has no rate limit of its own to impose on you. What it has is a set of obligations to the public services it reads, and running it makes them yours.
Written by the people who wrote the software, with no legal review of any kind. It is accurate about what the code does, which is the part they can vouch for. It is not a warranty that it satisfies any statute in any jurisdiction.
What you agree to when you run this
- Identify yourself to the sources. SEC EDGAR requires a declared User-Agent naming a real contact; requests without one are refused and, repeated, get an address blocked. This app declares one — check what it is on the source health page before you deploy.
- Respect the published rate limits. EDGAR permits ten requests a second. The app's caches exist substantially to stay under limits like it, and disabling them to "go faster" moves the consequence onto a public service funded by taxpayers.
- Do not redistribute what you are not licensed to. Fetching a source for your own analysis and republishing its content are different acts under most of these licences. The licences page has the terms per source.
- Do not present this software's output as regulated advice. See the disclosures page.
What this software will not help you do
The project's own brief says it, and it is worth repeating where a reader can find it: all intelligence gathering here relies exclusively on legal, publicly accessible sources — OSINT, FOIA disclosures, PACER dockets and SEC filings.
- No ingesting, simulating or linking to unauthorised or proprietary internal data of any institution.
- No mock data, synthetic placeholders or fabricated endpoints, anywhere, for any reason. When a source is unavailable the app says so and names it.
- No scraping of paid terminals, and no circumvention of any provider's access controls.
If you host this for other people
You become the operator, and every obligation above becomes yours on their behalf. Two things are worth setting before anyone else can reach it: a contact address, so a security report has somewhere to go, and the declared User-Agent, so a source that needs to reach you can.