Bulk earnings calls

Download every earnings-call transcript as CSV files. Each file holds one fiscal quarter of full transcripts for all companies. One route lists every file together with a signed download link. Add year and period to get one file.

Signed download URLs are valid for one hour after URL generation. Start each file download within this window. If a URL expires, request a new one.

Available files

Browse the files and download them from the browser. Downloads require a Professional or Enterprise plan. The route returns the same list with your API key.

Parameters

Response

Each item is one file with its size, its SHA-256 digest, and a signed url. Items are sorted newest fiscal year first, then quarter order. Without filters the list holds every file, so the page URLs are always null. The example response is cut after two files. The response is not cached because every link expires one hour after the request.

File contents

Each file is a gzip-compressed CSV with a header row. Rows are sorted by symbol. One row is one call. The transcript column holds the full text of the call with each speaker turn on its own line.

The first rows of the fiscal 2026 Q2 file, with the transcript text cut short:

id,symbol,fiscal_year,fiscal_quarter,date,transcript
ecall_wq5rd85raHX4Jm,AMEX:ACCS,2026,2,2026-08-11,"Forrest MacConnell: Welcome to ACCESS Newswire's Second Quarter 2026 Earnings Conference Call. ...
Brian Balbirnie: Thank you, Forrest, and good morning, everyone, ...
Steven Knerr: Thank you, Brian, and good morning, everyone. ..."

Transcripts contain line breaks and quotes, escaped as RFC 4180 requires. Use a real CSV parser. Do not split the file on line breaks.

Fiscal years run ahead of calendar years for some companies, so files for the next fiscal year can already exist. The route is the authoritative answer for what exists.

Download, verify, and decompress

Run these commands to download one quarter, verify the digest, and decompress the file. Replace YOUR_API_KEY with your key.

# 1. Get the link and the digest of one file
RESP=$(curl -s -H "Authorization: Bearer YOUR_API_KEY" \
  "https://api.roic.ai/v3.0.0/earnings-calls/bulk?year=2026&period=q2")
URL=$(echo "$RESP" | jq -r '.data[0].url')
SHA=$(echo "$RESP" | jq -r '.data[0].sha256')
 
# 2. Download the file before the link expires
curl -o roic-earnings-calls-2026-q2.csv.gz "$URL"
 
# 3. Verify the digest
echo "$SHA  roic-earnings-calls-2026-q2.csv.gz" | shasum -a 256 -c
 
# 4. Decompress and keep the archive for the examples below
gunzip -k roic-earnings-calls-2026-q2.csv.gz

To read the transcripts in Python, raise the field size limit first. A transcript can be longer than the default limit of the csv module.

import csv
import gzip
import sys
 
csv.field_size_limit(sys.maxsize)
 
with gzip.open("roic-earnings-calls-2026-q2.csv.gz", "rt", newline="") as csv_file:
    for row in csv.DictReader(csv_file):
        print(row["symbol"], row["date"], len(row["transcript"]))

Errors

A year without files returns an empty data array. The file host bulk.roic.ai blocks a link without a valid verify token or after expires_at. Call the route again for fresh links.

Related endpoints

• Bulk downloads overview
• List earnings calls to find calls for one company without transcript text
• Retrieve an earnings call as JSON speaker turns