MODEL CONTEXT PROTOCOL

Your agent.
The whole story.

Help a compatible agent find your work, understand the context and use tools within its permissions. Keep decisions visible, wherever the conversation happens.

Read the connection guide Try Tellenze

An illustrative connection model · motion can be paused

WHAT YOU CAN DO

A question can lead to a useful next step.

MCP is an open standard that connects AI applications to external systems. In Tellenze, compatible clients can use it to find context and workspace tools within their permissions.

About the open protocol

Examples of workspace toolsDiscover the live tools before use. Available actions depend on your deployment, enabled features and current permissions.

Find the work and its context

Search projects, tasks, comments and knowledge you can access. Filter work by project, stage, assignment, blocked state or due date. Then read its acceptance criteria, links, discussion and source context within your permissions.

search_workspace
list_work
get_task_context

Create work and record progress

Create tasks in a project or team, edit their details, assign work, add comments and move it through the configured workflow. Read the current version before editing and record evidence when completing work.

create_task
update_task
move_task
add_task_comment
relate_tasks

Coordinate Focus and milestones

Plan a Focus around a goal, link work to milestones across projects and follow progress. A milestone link does not give access to private work.

list_focus_groups
create_focus_group
list_milestones
link_milestone_task

Keep shared knowledge useful

Find and read project or workspace documents, check previous revisions and save a new document or an agreed edit. Saved changes keep the author and revision history.

list_documents
get_document
create_document
update_document

Plan your day and week

Read your Focus day, choose up to three visible work priorities and save a weekly intention with priorities in order. Personal planning uses your access and working-day preferences.

get_focus_day
set_focus_day
get_weekly_review
update_weekly_review

Turn Signals into reviewed work

Submit and discuss evidence, request a limited processing run and review suggestions. With the right access, explicitly accept or link work using current versions and guidance for the destination. Signals requires Starter or above.

submit_signal
list_signals
get_signal
manage_signal
process_signals

Read the guidance before acting

Find the Required Context that applies, read every part of its pinned revisions and acknowledge delivery before a protected change. This acknowledgement records that the guidance was delivered. It does not prove the work is correct.

resolve_required_context
get_required_context_revision
acknowledge_required_context

Learn from delivery and report usage

Read delivery insights you can access and approved workshop outcomes. Prepare private review drafts and report measured agent usage. Members must still explicitly publish or share reviews and approve workshop outcomes.

get_delivery_insights
get_retrospective_outcomes
get_refinement_outcomes
create_review
report_agent_usage

A USEFUL STARTING POINT

Three steps to get going.

Choose your workspace

Connect a compatible MCP client to the main workspace endpoint shown in Settings → MCP.

Authorize as yourself

Sign in and select the machine profile for this connection. Its tools use your current membership and access to source material.

Start with the context

Discover tools and read the relevant work. Review proposed changes before you approve an external action.

IN PRACTICE

From a conversation to a recorded outcome.

These examples show possible workflows. Replace sample projects, identifiers and versions with your workspace values. Follow the current setup guide and, for MCP, the schema returned by tool discovery.

Find and read

Understand before changing

“Find blocked work in SUPPORT and explain the next step for SUPPORT-42.”

search_tools({ "query": "list_work get_task_context" })

execute_tools({
  "calls": [{
    "name": "list_work",
    "arguments": {
      "project_key": "SUPPORT",
      "blocked": true,
      "limit": 10
    }
  }]
})

Read a selected result with a separate get_task_context call. Tool discovery returns the current schemas, which describe each tool’s fields. Use the returned cursors to get more results.

Create a task

Record a concrete next step

“Create a follow-up in SUPPORT with this context and acceptance criterion.”

execute_tools({
  "calls": [{
    "name": "create_task",
    "arguments": {
      "idempotency_key": "support-followup-1842",
      "project_key": "SUPPORT",
      "title": "Verify the checkout fix",
      "description_markdown": "Follow up on the support report.",
      "acceptance_criteria_markdown": "Checkout succeeds on mobile."
    }
  }]
})

Read the destination’s work types and guidance first. When a policy applies, read and acknowledge Required Context, then add required_context_receipt_id to the write.

Read and revise

Keep the knowledge current

“Read the incident runbook, propose the new step, then save the agreed revision.”

list_documents → get_document
Review the complete proposed Markdown
update_document({
  "idempotency_key": "runbook-update-1842",
  "document_id": "DOCUMENT_ID_FROM_READ",
  "current_version": 1,
  "body_markdown": "# Runbook\n\nThe complete agreed content."
})

This example uses an ID placeholder. Use the current version from the read response. The supplied body replaces the whole document. Protected edits need a current context receipt.

Task email Powerup

Prepare a prospect email for review

“Prepare an introduction on LEAD-42. Show me the sender, recipients and complete copy before sending.”

Discover the task email tools
list_task_email_connections → prepare_task_email
get_task_email → review exact saved fields
update_task_email → review again if edited
Human authorizes this exact email
send_task_email → inspect its receipt

Requires the task email tools to be deployed and an eligible Gmail or SMTP sender. Sending needs the reviewed current_version, review_fingerprint and confirmed: true. Preparation never sends.

Full setup guide

CLEAR PERMISSIONS

A connection carries your access.

Current permissions

A connected agent does not gain access to hidden projects, another member’s private drafts or private workshop material.

Review before action

Preparing an email or provider proposal sends nothing. You must approve the reviewed destination and current fields before sending or running an action.

Control the connection

Revoke it from MCP settings. A suspended membership, archived profile or change in source access can stop an action that was previously available.

A FEW USEFUL DETAILS

Know what to expect.

Will every assistant work?

Use a client that supports the workspace’s MCP transport, authorization and tool discovery. These control how it connects, signs in and finds tools. Support varies by client; follow the connection guide for setup.

Is the ChatGPT app generally available?

The dedicated ChatGPT connection remains a private pilot pending public approval. You can set up a direct connection separately with a compatible MCP client.

Can MCP send an email?

Task email tools can prepare, read, edit and explicitly submit a reviewed email when those tools are deployed and a Gmail or SMTP sender is enabled. Discover the live tools first and check the receipt after sending.

CHOOSE THE RIGHT CONNECTION

Different ways in. Shared work.

COME TAKE A LOOK

A little less busy.
A lot more together.

Curious? Join the waitlist and see how we welcome your team.

Tell us a little about your team. We’ll explain the next steps and invite you when we have capacity for your workspace.

Join waitlist Start with your email. No payment needed to join.