Stride Connect / Model Context Protocol

Run Stride work from your LLM.

Access lessons. Submit selected evidence. Retrieve the recorded result.

Your AI application runs the conversation. Stride handles access, assessment rules, and progress.

READ + WRITE / EXAMPLE TRAFFIC
Operation
Find accessible work
find_microlearnings
Response
Matching lessons + identifiers
ILLUSTRATIVE FLOW · NO LIVE REQUESTSConcurrent illustrative requests · Not live traffic.
Available internally at CGI. Marketplace listing coming soon.MCP · Authenticated access · Structured results
The protocol

What is MCP?

Model Context Protocol, or MCP, is an open standard that lets AI applications connect to external tools and data. Stride Connect uses it to make your learning work available inside Claude, ChatGPT and compatible command-line tools.

The model can request an operation. Your AI application sends it to Stride. Stride checks access, runs the supported operation and returns a structured result.

Current implementation

Learning tools. Interactive views.

Eight learner tools cover discovery, assessment and recorded results. Today’s prepared source adds interactive views and reviewed product help, with their release status shown below.

READ / 05

Connection status

Read your connection state and recent assessment summary.

get_my_status
Current work

List your current eligible Stride work.

list_current_work
Lesson search

Find accessible microlearnings beyond the current-work list.

find_microlearnings
Work details

Retrieve task context, available criteria and recorded feedback.

get_work_detail
Attempt status

Read the recorded assessment state and outcome.

get_attempt_status

WRITE / 03

Start an attempt

Create or recover a private attempt and receive its frozen rubric.

start_attempt
Attach evidence

Add a selected excerpt or supported reference to your attempt.

add_attempt_evidence
Submit an assessment

Submit criterion scores, reasons and evidence for validation.

submit_assessment

Built in source.

AWAITING VALIDATION AND RELEASE

The current implementation also includes the prepared interactive and product-help packages. These additions are not yet available in connected workspaces. Build, testing, content approval and host acceptance remain release gates.

Visual work selection

A work picker and lesson-detail cards for finding learning and inspecting its requirements inside the AI application.

Recorded-result cards

Display an attempt’s recorded outcome and feedback. The first package adds read-only views and preserves the existing assessment tools.

Shared interactive interface

Prepared for Claude and ChatGPT, with text fallback, versioned UI resources and validated presentation contracts. Packaging and acceptance in both hosts are pending.

Scoped, controlled rendering

Resource authorization, default-off organization rollout, safe rendering and stale-card handling are included in the prepared source.

Reviewed product help

A separate shared help service and two proposed MCP search/article adapters are prepared. Tool registration and content approval remain on hold. They are not additional active tools yet.

Writes affect your own eligible assessment work. Practice does not submit an assessment.

Technical contracts ↗
The data boundary

Selected evidence. Not the entire chat.

You choose what to share. The connector sends that evidence and its assessment to Stride.

Read the access and data model ↗
SUBMISSION BOUNDARYSELECTED CONTENT ONLY
IN THE LLM APPLICATION
Conversation context
Exploration and practice
Other messages
Not included in the submission
✓ Selected work

A chosen excerpt or supported reference.

EXPLICIT SUBMISSION
TO STRIDE

Assessment payload

Evidence
Selected work or reference
Assessment
Criterion scores + reasons
← STRIDE RETURNSRecorded outcome + feedback
Connection setup

Connect your AI application.

Available internally at CGI. Your organization supplies the connection details and controls workspace access. Use the instructions for a client your organization has enabled. Contact Stride for current availability and connection details.

  1. Add the connection.

    Use the Stride connector in your approved workspace, or your organization’s custom-connector instructions.

  2. Sign in to Stride.

    Use your own account and review the requested access.

  3. Request your work.

    Ask for your current Stride work or a specific microlearning.

Get connection details

Beyond the learner workflow.

Interactive work picker, lesson detail and result cards are in development. Product help, content discovery, reviewed contributions and role insights follow separate release stages.

Explore Coming Soon →
CGI Stride Connect