MCP (Model Context Protocol)
LetsLigo exposes an MCP (Model Context Protocol) server that allows AI assistants (Claude, ChatGPT, Cursor, etc.) to access your contacts, tasks, and CRM pipeline directly from their interface.
MCP is an open-source protocol that standardizes the connection of AI applications to external systems. By connecting LetsLigo via MCP, your assistant can list your contacts, search for a contact, view your tasks or pipeline states — all in natural language.
MCP server URL
https://www.letsligo.com/_mcpAuthentication
Authentication uses the same API token as the REST API. Create a token in Settings > Developer of your workspace, then pass it in the Authorization: Bearer YOUR_TOKEN header when connecting to the MCP server. Add X-Workspace-Slug if you have multiple workspaces.
Direct MCP server access and API token creation are reserved for the Enterprise plan.
Available tools
The LetsLigo MCP server exposes the following tools, usable by AI assistants:
contact_list
Lists the latest contacts in the workspace.
contact_search
Searches contacts by name, company, email or phone.
contact_get
Retrieves contact details by ID (UUID).
task_list
Lists pending tasks in the workspace (optional: yours only).
task_get
Retrieves task details by ID (UUID).
pipeline_states_list
Lists CRM pipeline states (New contact, Qualified, Client, etc.).
company_list
Lists companies in the workspace. A company can be shared by multiple contacts.
company_get
Retrieves company details by ID. Use contact_get to see companyId for a contact's company.
company_update
Updates company information. Use company_get or contact_get to obtain the companyId. Always ask for confirmation.
contact_create
Creates a new contact. Always ask for user confirmation before calling.
contact_create_prepare / contact_create_confirm
2-step flow: prepare returns a summary and token, confirm executes after validation.
contact_update
Updates an existing contact by ID. Always ask for confirmation before calling.
contact_update_prepare / contact_update_confirm
2-step flow: prepare returns a summary and token, confirm executes after validation.
contact_update_pipeline_state
Changes the CRM pipeline state of a contact. Use pipeline_states_list for state IDs.
contact_note_add
Adds a note to a contact profile. Use contact_search or contact_get to get the contactId. Use when the user asks to add a note, comment, or information to an existing contact.
contact_note_delete
Deletes a note from a contact. Use contact_get to see notes with their noteId. When user asks to delete a note, use this tool — NEVER create a note with 'Delete' or similar as content. Always ask for confirmation before deleting.
task_create_prepare / task_create_confirm
2-step flow: task_create_prepare returns a summary and token. When user confirms ('yes', 'OK', 'confirm'), call task_create_confirm IMMEDIATELY with that token. Never respond with a greeting when user is confirming.
task_complete
Marks a task as completed.
email_prepare
Generates an email for a contact with full context (meeting, enrichment, workspace refs). Provide an instruction (e.g. 'Write a follow-up email'). Returns subject, body, sessionId for iterations.
email_refine
Refines the generated email with an instruction (e.g. 'Shorten paragraph 2'). Requires sessionId from email_prepare. Supports iterations like the wizard.
email_send_prepare / email_send_confirm
2-step flow: prepare returns summary and token, confirm sends the email after validation.