Dr. Priya N. is an associate professor of finance whose research focuses on management communication patterns and how they predict firm outcomes. She's been using earnings call transcripts in her work for years, but until recently, data access was the bottleneck. Commercial databases were priced for institutional investors, not university budgets. We talked about how a $24.99 API plan made her research affordable, and why her undergraduate students now read transcripts for class without anyone paying a cent.
Interview recorded on August 14, 2026. Plans and prices mentioned are the ones the interviewee was on at the time; current plans are on the pricing page.
Dr. N., how does an academic use an earnings call API?
I study language patterns in how management communicates, specifically whether hedging language in CFO remarks predicts subsequent earnings restatements. Earnings calls are the ideal dataset for this: spontaneous, unscripted, repeated quarterly, and rich in psychological signal. The problem has always been data access. Hand-collecting transcripts is impossible at scale. Commercial databases cost thousands per year. For most academics, that's simply not available.
Earningscalls.dev changed that. The Pro API at $24.99 fits within a departmental discretionary budget. I can pull speaker-segmented transcripts for a defined universe of companies across five years. The roles are pre-tagged, which matters a lot for behavioral research. I have to know I'm analyzing the CFO and not the CEO or an analyst.
What's your current project?
I'm examining whether linguistic markers of hedging in CFO prepared remarks predict earnings restatements. I need clean speaker attribution, because the CFO's words are the whole study. With the API, I filter on executive role and use the speaker name field to isolate the CFO. The structured format means my research assistant can reproduce the entire dataset with the same API calls. Reproducibility is basic for academic work, and with manual collection it never really worked.
How does the free reading tier fit into your teaching?
My undergraduate students read transcripts for a class assignment on earnings communication. Before, I'd have to put PDFs on the course management system, which raised copyright questions and formatting issues. Now I give them a link to earningscalls.dev, and they can search and read without an account. No login, no paywall, no friction. For the field, that open access matters more than the API itself. Students at teaching-focused schools, not only at research universities, can work with primary sources.
What are the limitations for academic use?
The 2020 start date is the main constraint. Five years is enough for a working paper but not for a book-length study of, say, the 2008 financial crisis. I've adjusted my research questions to fit the available data. Also, coverage skews toward US and European large-caps. For a global comparative study, I'd need to verify coverage before committing. These are known constraints, and the platform is transparent about them.
How does the cost compare to alternatives?
The alternatives I looked at started at several thousand dollars per year for academic access. That's simply not available to most researchers. At $24.99 per month, this is accessible to PhD students, adjuncts, and researchers at teaching-focused institutions. The free tier matters even more, because my students can read transcripts without anyone paying anything. I've recommended it to colleagues in my department, and two of them are now using it for separate projects.
What about the MCP server? Have you experimented with it?
I've used it for exploratory analysis. Asking Claude "what did this CFO say about litigation reserves over the last four quarters" is a quick way to surface patterns before formal coding. It doesn't replace rigorous analysis, but it's a good first pass. For a researcher who isn't a programmer, the MCP interface lowers the barrier significantly.
Any advice for other academics considering this?
Start with the free tier on RapidAPI to test whether the coverage meets your needs. Fifty requests a month is enough to validate your universe. Then move to Pro for the actual research. The documentation is clear enough that researchers without engineering support can get started. That's rare for financial data APIs, and most academics don't have a data team.
What's been the most surprising finding from your work so far?
The signal is stronger in the Q&A than in prepared remarks. That makes sense. Prepared remarks are scripted and polished, the Q&A is unscripted and shows the cognitive load. The API's speaker segmentation is what allows me to separate the two and test that hypothesis rigorously. Without the structure, I'd be guessing.
Want to try the same workflows? You can search a whole sector, pull speaker-tagged transcripts as JSON or connect Claude through the MCP server. Plans and coverage are on the pricing page, and the full reference is in the API docs.