Lyearn MCP uses the Model Context Protocol, a standard for connecting AI apps to external tools and information. This guide covers connecting Lyearn to Claude and using the available capabilities.
Availability
Lyearn MCP is included in every paid Lyearn plan at no additional charge. Users in a paid customer organisation can connect immediately without requesting separate activation from Lyearn.
Each person connects with their own Lyearn account. Existing Lyearn permissions determine the tools and information available to that person. Your AI app may also require a workspace administrator to enable the connector.
What you can do
| Area | Examples |
|---|---|
| Company knowledge | Find recorded information about products, strategy, competitors, policies, and decisions. |
| People and organisation | Explore teams, roles, reporting relationships, locations, and open positions. |
| Learning | Find learning resources, inspect course details, search searchable material, and get relevant recommendations. |
| Learning activity | Review progress, overdue learning, certificates, badges, and points within your access permissions. |
| Analytics and productivity | Explore available reports, productivity metrics, benchmarks, and MGSA scores. |
| Organisation proposals | Save proposed structural changes for human review and publication. |
The available information depends on the features configured and records maintained by your organisation.
Connect to Claude
Use Claude on the web or desktop for setup. Have your Lyearn account and work email ready.
Server URL: https://api.use1.lyearn.com/productivity/mcp
Add the connector
| Claude account | Where to start |
|---|---|
| Individual account | Customize → Connectors; add a custom connector. |
| Team or Enterprise workspace | An authorised owner or administrator adds a custom web connector from Organization settings → Connectors. Members then connect individually. |
Name the connector Lyearn and enter the server URL. Select OAuth authentication and Use Claude’s published identity where offered, then finish adding it. Standard setup does not require a client ID or secret.
For Claude’s current interface and account requirements, see its custom connector setup guide.
If Lyearn is available in your Connectors Directory, you can select that listing and follow its connection prompts instead.
Sign in to Lyearn
Select Lyearn in Claude’s connectors and start the connection.
Enter your work email or workspace when prompted.
Sign in to your Lyearn workspace using your usual password or your organisation’s single sign-on.
Review the requested permissions and authorise the connection.
Enable Lyearn for the conversation.
Some organisations use a connection configured by their administrator. If you are asked for a client ID or secret, contact support@lyearn.com for the appropriate setup instructions.
Check the connection
Ask:
Use Lyearn to show the learning resources I can access.
A successful response should retrieve the accessible library or explain that it contains no matching items. You can also ask which analytics reports are available. A permission refusal means that capability is restricted for your account; it does not necessarily mean the connection failed.
Use Lyearn through conversation
Name the team, learning item, person, or time period you want to work with. You can ask the AI app to show supporting records, explain the access scope, and point out missing data.
Find and understand learning resources
Start with:
Find learning resources about handling customer objections. Show their objectives, duration, and relevant skills where available.
Choose a course or article from the results, then ask:
Summarise the key techniques taught in this item and show the passages supporting the summary.
The result should distinguish the item’s recorded description from what its searchable material teaches. Searching within material and grounded summaries require searchable course or article content.
Review team learning
Ask:
Show overdue learning for my team, including the learning items, enrolled counts, and overdue counts. State whose learning activity I can access.
Then narrow the question:
For the first item, show the learners who are overdue and their due dates.
Team views require the relevant team-learning permission. Company-wide views require broader access.
Understand productivity scores
Ask:
Explain my latest available MGSA scores, the underlying metrics, and any gaps in the evidence.
MGSA covers mindset, goals, skills, and work activity. Scores depend on the productivity models and evidence available in your organisation. Missing scores are unavailable information; they are not scores of zero.
You can follow up with:
Recommend accessible learning resources relevant to the skills I need to improve. Explain why each resource was selected.
Explore reports
Ask:
List the available learning reports. Using a suitable report, show completions by department for last month. State the date range and reporting scope.
The AI app can use available report columns, aggregations, and filters. Productivity analyses must come from the analyses defined in an enabled model.
Review your organisation and prepare changes
Start with a read:
Review the Sales reporting structure for inconsistencies. Explain the supporting records and show proposed corrections before saving anything.
After reviewing the suggestions, you can ask the AI app to save the agreed changes as a proposal or named draft. This creates a reviewable record in Lyearn.
An administrator reviews individual proposals, accepts or dismisses them, and publishes accepted changes through Lyearn. Saving a proposal does not itself change the published organisation.
Permissions and actions
Connecting gives the AI app access through the permissions granted to your connection and the permissions of your Lyearn account. Adding a connector does not grant a user additional Lyearn permissions.
| Action | What happens |
|---|---|
| Search, retrieve, summarise, or analyse | The connector returns information through available read tools. |
| Save an organisation proposal or draft | A reviewable record is created or updated in Lyearn. |
| Publish proposed organisational changes | A human administrator reviews and publishes accepted changes in Lyearn. |
Your own learning information and information about other learners have different access requirements. Team-learning access follows your reporting scope; company-wide access requires the relevant broader permission.
Review requested permissions when connecting. For control over tools in Claude, see Claude’s connector management guidance.
Data handling
Information shared with the AI app
Lyearn returns information relevant to the query being answered, within the connection’s access permissions. Depending on your request, that can include company knowledge, people and organisation records, learning content and activity, reports, or productivity information.
Information returned to the AI app is processed under that provider’s terms and privacy settings. Lyearn’s deletion process applies to information held by Lyearn; information already returned to another provider must be managed through that provider.
Logging and temporary storage
Lyearn uses operational request logging, telemetry, and a temporary replay cache to operate the connector and handle repeated write requests.
| Category | What it contains | Retention |
|---|---|---|
| Tool telemetry | Tool names, categories, outcomes, timings, and trace identifiers. Telemetry excludes user identities, tool arguments, and tool results. | 90 days. |
| Temporary write replay cache | Results from certain write operations, used to handle repeated requests. Results may contain workspace information. | Successful results are retained for 10 minutes. |
Operational request logs are maintained separately from the telemetry described above.
AI model training
Lyearn does not use information accessed through MCP to train AI models. The connected AI provider’s handling of information is governed by that provider’s applicable terms and account settings.
Read the Lyearn Privacy Policy for Lyearn’s general data practices.
Disconnect and request deletion
You can manage and revoke your connection from the profile menu in Lyearn. An authorised integrations administrator can manage connections from Settings → Manage → AI app connections.
You can also disconnect or remove Lyearn from Claude’s connector settings. For the current controls, see Claude’s connection management instructions.
Disconnecting stops the connector from acting as that person. It does not delete information Lyearn already holds.
To request deletion, email support@lyearn.com. Lyearn deletes the requested Lyearn-held data, including associated connector records and logs, within 24 hours of receiving the request.
Limits and interpreting answers
Search and report results can be limited or paginated. Ask whether an answer covers the full population.
A summary reflects the available source material. Missing or unsearchable material can limit the answer.
Organisational headcount can count filled positions rather than every person on record.
Scores, metrics, and recorded history can have different update times. Check the relevant date and evidence coverage.
Rate and request-size limits apply. Narrow large requests by team, learning item, or time period.
If saving a proposal times out, check the existing draft or review queue before submitting it again.
For a grouped list of tools and important capability limits, see the Lyearn MCP capability reference.
Troubleshooting
| Issue | What to do |
|---|---|
| Lyearn does not appear in the directory | Add it as a custom connector using the server URL in this guide. |
| Your Claude workspace blocks the connector | Ask the workspace owner or authorised administrator to enable it. |
| Sign-in fails | Check that you are using the correct Lyearn workspace and account, then reconnect. |
| A tool is unavailable or access is denied | Ask your Lyearn administrator to review your permissions. |
| A person or learning item cannot be found | Use a more specific name or title and confirm the record exists in your workspace. |
| A course cannot be searched or summarised | Ask for its recorded description and structure; its material may not be searchable. |
| A report or metric is unavailable | Ask which reports, models, and analyses are available before narrowing your request. |
| A request is rate-limited or times out | Wait for any retry interval shown, reduce the scope, and check whether a write completed before retrying it. |
| A connection has expired or been revoked | Reconnect and complete the sign-in and consent flow again. |
Support
Email support@lyearn.com for connection problems, tool access questions, or deletion requests.
For troubleshooting, include your work email, workspace, AI app, the approximate time of the issue, the action you attempted, and the error or correlation ID if available. Keep passwords, access tokens, and client secrets out of the message.
You can also visit the Lyearn Help Center or check Lyearn service status.
