Both routes below use the same CSV file and the same rules as Organisation → Import people from CSV: name,email,jobRole,department,groups,managerEmail,access,startDate,active Ann Lee,ann.lee@example.com,Head Chef,Kitchen,night;day,,manager,2025-04-01, Bo Smith,bo.smith@example.com,Porter,Kitchen,night,ann.lee@example.com,employee,2026-01-12, name and email are required; the other columns are optional and can be in any order. People are matched on email, ignoring capitals. A new email adds a person; a known email updates only the columns the row fills in, so a blank cell keeps what is there. Nobody is deleted. active set to false (also no, 0 or inactive) marks a leaver and closes their open training. A later row with true brings them back. managerEmail is matched after every row is read, so row order does not matter. At most 5,000 rows and 8 MB per file. Every run is recorded in the audit log. ### By API Send the file to POST /api/lms/members/import with an API key from the organisation owner's account (Dashboard → Settings → API keys; see API overview): curl -X POST https://www.thediscourse.ai/api/lms/members/import \ -H "Authorization: Bearer dsk_…" -H "Content-Type: text/csv" \ --data-binary @people.csv The body can be: Content-Type: text/csv with the file itself, Content-Type: application/json with {"csv": "…"}, or Content-Type: application/json with {"rows": [{"name": "…", "email": "…"}]}, where groups may be a list. The answer gives the counts and any problem rows: {"created": 1, "updated": 2, "deactivated": 1, "errors": [{"line": 4, "error": "a valid email is required"}]} If the owner's account also belongs to another organisation, add the header x-lms-org with your organisation id so the import cannot land in the wrong one. An API key can reach this import, the Power BI feeds below and the course routes only. ### By SFTP drop folder For HR systems that can only export a file on a schedule, we set up an SFTP account that can see only your organisation's folder. Email admin@thediscourse.ai with your SSH public key; we send back the host, port and user name. Upload as people.csv.tmp and rename it to .csv when the upload has finished, so a half-written file is never read. Files are imported once a day, in name order, so name them by date (people-2026-11-05.csv). After the import, each file moves to processed/ with a .result.json beside it giving the counts and any errors. A file that cannot be imported is moved the same way with its errors and is not retried: fix it and upload it again. Clear out processed/ as your retention policy says; it holds personal data. Imported people are not emailed until they sign in and choose the organisation, as on Set up your organisation.
Each person has a private calendar feed of their own due dates and accreditation expiry dates, as all-day events ("Due: course", "Expires: accreditation"). Due dates include any extra days from a reasonable adjustment. They add it from My learning → Add to Outlook/Teams calendar; the address is also in every reminder email. The feed updates itself. The address is private to the person and only ever shows their own items; it stops working if they are marked inactive.
Managers' daily reminder digests can also be posted to a Teams channel: In Teams, on the channel, create an incoming webhook (under the channel's Workflows or Connectors) and copy its address. It must be an https address on webhook.office.com or logic.azure.com. Paste it into Organisation → Settings → Microsoft Teams webhook (optional) and press Save settings. Clear the field to stop. Posts go out only when reminders are on (see Compliance and reminders). Only administrators can see the saved address.
Organisation admins can report on their training data in Power BI (or Excel, or any tool that reads JSON from a web address). There are three feeds: To connect: Create an API key on the organisation admin's account (Dashboard → Settings → API keys). Only admins can read the feeds. In Power BI Desktop choose Get Data → Web, then Advanced. Put a feed address in URL parts. Under HTTP request header parameters choose Authorization and enter Bearer dsk_… with your key. When asked how to sign in, choose Anonymous: the key in the header is the credential. Choose To Table, expand the records, and set the date columns to Date/Time. Repeat for each feed you want. After you publish the report, set Scheduled refresh on the semantic model in the Power BI service; the credentials stay Anonymous. Every read is recorded in the audit log as "Power BI read". Dates are ISO 8601 in UTC and empty values are null. If the account belongs to more than one organisation, add the header x-lms-org with your organisation id. Treat the key like a password, since it can read your whole organisation's training data; delete it under API keys if it leaks.
These settings go in the app's .env and need a restart:
Response and resolution times are on Service levels. Email admin@thediscourse.ai; put "Critical" at the start of the subject for an outage at any hour.
Back to Set up your organisation, or on to the API overview.
Pricing · What the training record shows · Set up learning management · FAQ · Book a 20-minute walkthrough