Skip to content

SIBiLS documents and annotations fetch API


This API allows to retrieve annotated contents from a given collection. The input is a set of docids (up to 1,000 per request). The output is a set of parsed and annotated documents, in json and/or BioC formats.

API endpoint


Mandatory input: one collection (&col=), amongst "medline", "pmc", "plazi" and "suppdata" Mandatory input: the list of requested document ids (&ids=). Docids are separated by commas. Optional input: the output format (&format=) amongst "json" (default) or "BioC" (to come)

Example: fetch two PMIDs from MEDLINE,25190367&col=medline

Code sample

python script for demonstrating POST calls to the API

import requests # not installed in default Python
import json

# PMIDs to fetch (limited to 1000)
pmids = "14691011,25190367"
collection = "medline"

# call with POST
url_API = ""
my_params = {"ids": pmids ; "col": collections} # parameters dictionary
r = = url_API, data = my_params)

# get response and print in output
response = r.text
with open("SIBiLS_fetch.json","w",encoding="utf-8") as file:


Output is json and/OR BiOC BioC formatted :

    [Array of documents corresponding to the query]
            [Array of PubMed like informations about the document]
            [Array of passages splitted from the document]
                [Array of informations relative to the passage]
                {"type": nature of the passage (e.g. "title", "abstract")
            {"text": the passage
                [Array of annotations relative to the passage]
                {"text": concept annotated
                    [Array of informations about each annotation]
                    {"type": type of entity (text) matched
                    {"source": terminology used
                    {"source_id": ID from the terminology
                    {"searchable_id": code(terminology)+";"+ source_id
                    {"prefered_term": prefered_term extracted from the terminology
                    {"sentence": sentence where the text is found
                    {"offset": start position of the text in the sentence
                    {"length": length of the text annotated