Skip to main content
Spott provides a hosted MCP (Model Context Protocol) server at https://mcp.spott.io/mcp. It exposes a set of tools that AI assistants can call to work in Spott on your behalf: searching records, creating candidates, moving applications, scheduling interviews, sending emails, logging calls, and more.

​
What it’s for

  • Run your day-to-day recruiting workflow from a chat conversation.
  • Build assistant-driven automations that read and write Spott data.
  • Connect Spott into multi-step agentic flows alongside other connectors.

​
Set it up

The quickest paths are our step-by-step guides:

Use Spott in Claude

Connect Spott in Claude apps, or add it as a custom connector.

Use Spott in ChatGPT

Install the Spott plugin from ChatGPT’s plugin directory, or add it with the MCP server URL.

​
Other AI assistants

Spott is not limited to Claude and ChatGPT. Other AI assistants, such as Copilot, can connect too, as long as your plan for that tool supports custom MCP servers. Whether it does depends on the tool and the plan, so check with its provider. There is no step-by-step guide for these tools yet, and the setup happens on the tool’s side, so follow its own instructions for adding a custom MCP server. The general flow is:
  1. Add a custom MCP server (some tools call it a custom connector) and enter the server URL by hand: https://mcp.spott.io/mcp.
  2. Connect. The tool redirects you to a Spott OAuth page.
  3. Approve access. The assistant can now make authenticated calls on your behalf.

​
Authentication: OAuth, per user

Unlike the REST API, which uses API keys, the MCP server authenticates with OAuth:
  • Each user connecting completes their own OAuth flow; access is scoped to that user.
  • Actions the assistant performs are attributed to the connected user.
  • No keys to paste or rotate, and access can be revoked from the Spott side.

​
The connection needs a permission of its own

Connecting an assistant is gated by a role permission, granted under Settings → Security. It is separate from the record permissions the connector inherits, so a user with full access to candidates and jobs can still be unable to connect. Without it the failure is confusing rather than obvious: the sign-in completes normally, the connector reports itself as connected, and then every request comes back as MCP_ACCESS_DENIED. If that is what you are seeing, ask an admin to add the permission to your role. See managing access.

​
What the assistant can do

The tool set grows over time and reflects the live capability surface: searching and looking up records, creating and updating candidates, contacts and companies, managing applications and pipelines, notes, tasks, emails, calendar invites, placements, and more. The assistant can also read the transcript of a call or meeting, list your meetings, and read the contents of an attachment, such as a CV or a job spec, so it can answer from what was said and written rather than only from the record’s fields. After connecting, ask your assistant what Spott tools it has available for the current list.

​
MCP vs the API vs webhooks

  • MCP: AI-assistant-driven interactions and agentic flows.
  • REST API: traditional integrations and programmatic data access at scale.
  • Webhooks: event-driven notifications from Spott to your systems.
The three are complementary and can be combined.