Announcing our new PACER Fetch APIs

Until today, if you wanted to add something to the RECAP Archive on CourtListener, you had only one option: Use the RECAP Extensions to purchase the item from PACER, and let the extensions do the upload on your behalf. While that works well, many of our users — especially those that use our APIs — have asked for something more. Could we provide an API for them to more easily get PDFs and dockets from PACER? As of today, with the support of the employment law firm Jet.law, the answer is finally yes. Starting today, we have a new free API released in Beta that anybody can use to get dockets and PDFs from PACER and add them to our website, APIs, and replication systems.

This is a tool for the techies, so here's how it works. Let's say you want a docket from PACER. Get it with:

curl -X POST \
    # Type 1 is for dockets  
    --data 'request_type=1' \
    # The docket number and court you want
    --data 'docket_number=5:16-cv-00432' \
    --data 'court=okwd' \
    # Your PACER and Courtlistener credentials
    --data 'pacer_username=john_w_powell' \
    --data 'pacer_password=coloradoadventures' \
    --header 'Authorization: Token <your-token-here>' \
    https://www.courtlistener.com/api/rest/v3/recap-fetch/

That'll add a request to our fetch queue, and we'll go get it for you as soon as possible.

PDFs are similar, but they can only be done by CourtListener ID for now. This ensures that we don't get a PDF before we have the docket it belongs to. Here's how you'd get a PDF with the CourtListener ID of 112:

curl -X POST \
    # Type 2 is for PDFs
    --data 'request_type=2' \
    # The ID of the item you want
    --data 'recap_document=112' \
    # Your PACER and CourtListener credentials
    --data 'pacer_username=john_w_powell' \
    --data 'pacer_password=coloradoadventures' \
    --header 'Authorization: Token <your-token-here>' \
    https://www.courtlistener.com/api/rest/v3/recap-fetch/

Basically, it's the same as above, but with the request_type changed, and with the ID provided instead of the docket number and court. We take it from there.

There are several variations and approaches that you can learn about in our documentation for the feature. There are a couple questions we anticipate though.

1. What's the big advantage here?

Well, for one thing, if you want PACER content in JSON format, now you can easily get it. You don't have to set up crawling infrastructure. You don't have to write crawlers (or use ours). You don't really have to do much beyond creating a PACER account and sending us API requests. Just ask for what you want, then crawl the response when it appears in our APIs. We have literally hundreds of regression tests on our parsers. Trust us that you don't want to be in the crawling business if you can avoid it.

2. Will this always be free?

That's our goal. Our thinking is that if you want to use our tools (like the RECAP Extensions or these APIs) to add content to the public commons and the RECAP Archive, that benefits everybody and is so well-aligned with our mission that we shouldn't stand in your way.

3. How secure is this? I give you my password???

Unfortunately the only way to make this work is for you to share your password with us. But:

1. Your password is fully encrypted in transit to us.

2. Once we get your password, we immediately use it to log into PACER and get cookies for your account. We then throw away your credentials, keeping only the cookies. We never store or log your credentials, so they're typically only on our system for a few seconds. We don't have any columns in our database for your PACER credentials.

3. We save your cookies for one hour with an automatic expiration. After that time, we will need your username and password again.

4. We are leading the legal technology industry by having a vulnerability disclosure policy and bug bounty program.

For more on this topic, please see the security documentation on our website.

4. What about performance? How fast is this?

We aim to get each download done in a second or two, with many concurrent requests possible simultaneously. A request may take longer depending on a few things though:

• Some PACER websites are faster than others. Typically, more populous jurisdictions seem to have better PACER servers. Business hours are worse than after hours, etc.
• Some PACER dockets, particularly large ones or ones with hundreds of parties, can be exceedingly slow. This API lets you download truly huge dockets that can take a very long time for PACER to generate and serve.
• We extract text from every PDF we get. If OCR is needed, that usually takes about one second per page, though the item will be in our database while OCR is being computed.

5. How do I get started? Do I need to sign up?

Here's the documentation:

Start there. No need to tell us you're using the API, but feel free to get in touch if you have questions or if things aren't working like you'd hoped. This will be an evolving project, and we welcome your feedback and feature requests.

We can't want to see how people use these APIs.