Seeking advice on public research for 2k cases.

[ziboobiz723]

Hi CourtListener team,

I am building a public-data legal research workflow for a client who has around 2,000 source cases cited in his books. Each year, he wants to identify any new public cases that cite those source cases, limited to a requested annual update period, for example cases filed in 2026.

The intended workflow is:

Given a spreadsheet of source cases/citations, resolve each source case in CourtListener, find public CourtListener cases that cite it, filter those citing cases to the requested filing period, and produce a review spreadsheet for manual legal review.

My current REST API plan is:

1. Use citation-lookup to resolve each source citation to a CourtListener opinion/cluster.
2. Use opinions-cited with cited_opinion=<source_opinion_id>.
3. Filter by citing_opinion__cluster__date_filed__range for the annual period.
4. Fetch opinion/cluster metadata for candidate citing cases when needed.
5. Cache source resolutions and citing edges locally so future annual runs use fewer API calls.

In testing, one source case usually takes around 2 API calls if no new citing cases are found, and around 5-8 calls if one or more citing cases are found and metadata hydration is needed.

A small 27-case sample run worked, but before scaling to around 2,000 source cases, I want to make sure I am using CourtListener in the recommended and respectful way.

Could you advise on the best official approach for this use case?

Specific questions:

1. Is opinions-cited with citing_opinion__cluster__date_filed__range the right way to find citing cases filed within a specific annual period?

2. For around 2,000 source cases, would you recommend REST API usage with conservative pacing and local caching, or would you recommend building from the bulk data instead (bulk data download take hours and require huge disk space)?

3. If using the REST API, what access tier or rate limit would you suggest for this kind of annual batch run?

4. Are there any completeness limitations I should clearly disclose to the client when using CourtListener citation graph data for this purpose?

Thanks, I want to make sure the workflow is useful for the client while staying respectful of CourtListener infrastructure.