Docebo

The Docebo source ingests the course catalog of a Docebo LMS instance through the Learn REST API. Each course becomes an HTML document and carries the course description together with a curriculum outline of the course's training materials. Questions such as "which training covers this topic?" are answerable from the course record alone. Category, language, and skills become filterable metadata.

What is ingested

The source ingests courses, not the training materials inside them.

A course's materials — SCORM and xAPI packages, videos, quizzes, and HTML pages — are not downloaded; Docebo does not serve their contents to a server-to-server integration. Each material's title, type, and description are rendered into the course document as a Course contents outline.

Docebo's catalogs are not ingested, and catalog membership is not available as document metadata.

Authentication

auth is required. Before you create the pipeline, create an OAuth 2.0 application in the Docebo administration interface under Settings → API and SSO, and make sure the application is permitted to read courses. An application without that permission receives 401 Unauthorized from the course endpoints even though its token is valid.

Supply the credential as a RemoteAuth credential.

auth.type	Fields	Resulting header
oauth_client_credentials	client_id, client_secret, token_endpoint, and optionally scopes	A fetched Authorization: Bearer <token>
bearer	token	Authorization: Bearer <token>
header	header, value	The value sent verbatim under the named header

Docebo issues OAuth 2.0 client credentials for server-to-server access, so oauth_client_credentials is the usual choice. Set token_endpoint to {base_url}/oauth2/token and scopes to ["api"]. Use bearer for an instance that issues a standing API token instead. Access tokens are short-lived (Docebo's expire after an hour); the platform acquires and refreshes them for you. Credential secrets are encrypted at rest and never returned in responses.

Configuration

SOURCE FIELD (DOCEBO)

Code example with json syntax.json

1

Fields

Field	Required	Description
type	Yes	docebo.
base_url	Yes	Base URL of the Docebo instance, for example https://example.docebosaas.com. The Learn API paths are appended automatically and must not be included.
auth	Yes	Authentication for the instance. See Authentication.
published_only	No	Restricts ingestion to published courses, skipping any course Docebo reports as unpublished. Defaults to true.
include_languages	No	Docebo language codes to ingest, for example english or japanese, matched case-insensitively. A catalog commonly carries the same course once per translation, so this selects which translations are ingested. When omitted, every language is ingested.
exclude_languages	No	Docebo language codes whose courses are excluded, even when they also match include_languages. When omitted, nothing is excluded.
include_categories	No	Docebo category names to ingest, matched case-insensitively. Use this when the catalog mixes public training with internal or partner-only branches. A course with no category is skipped whenever this is set, so an uncategorized course is never silently treated as public. When omitted, every category is ingested.
exclude_categories	No	Docebo category names whose courses are excluded, even when they also match include_categories. When omitted, nothing is excluded.

How records are fetched

Each run pages through /learn/v1/courses, then reads the full record for each selected course from /learn/v1/courses/{id}. The course document contains, in order: the course name as a heading, the short description and description (embedded as the HTML fragments Docebo returns), the Course contents outline, Skills, Instructors, and any course custom fields the course filled in.

The outline is built from the course's training-material folders, listing each material's title, type, and description. A course with no training materials — an instructor-led classroom course, for example — gets no outline section.

Source metadata

Each record carries source metadata that the connector resolves at fetch time.

system_metadata:

Key	Description
title	The course name.
url	The learner-facing course page, {base_url}/learn/courses/{id}/{slug}.
code	The course code.
uid_course	Docebo's unique course identifier.
updated_time	The course's modified time, from Docebo's date_last_updated. Only the course listing reports this field, so it is absent on a record fetched by id, such as one reprocessed from a dead letter.

user_metadata:

Key	Description
course_id	The numeric course id.
category	The course's category name.
course_type	The Docebo course type, for example elearning or classroom.
language, language_label	The course's language code and display name.
credits	The course's credit value.
duration_seconds	The course's duration.
is_published	Whether Docebo reports the course as published.
skills	Comma-separated skill names associated with the course.
training_material_tags	Comma-separated distinct tags across the course's training materials.

The connector marks every record org-wide readable and applies no access policy of its own. Docebo enrollment is a learner-facing concept rather than a document grant, so it is not translated into record ACLs; narrow what is ingested with published_only and the include/exclude filters instead.

Incremental sync

The Docebo course listing exposes no server-side modified-since bound and no sort control, so every run re-enumerates the whole catalog and applies the change window client-side against each course's date_last_updated. Records therefore do not arrive in modified-time order, and per-record deduplication absorbs the courses that are re-listed but unchanged.

Docebo's timestamps carry no timezone, so both ends of the change window are widened by the widest possible timezone offset before they are applied. This is what causes the re-listing described above.

A course whose date_last_updated is missing or unparseable is ingested on every run rather than skipped, since without a timestamp its change state is unknowable.

note

A course's date_last_updated reflects edits to the course record. If your instance does not update it when a training material inside the course changes, schedule a periodic full refresh so curriculum edits are picked up.