Home → Release Notes → Release Notes → HelpSpot 5.8.0

1.2. HelpSpot 5.8.0

Updated: Jun 9 2026, 03:12 PM

Release Date: June 9th 2026

Featured: MCP Server

HelpSpot now ships with a built-in Model Context Protocol server, turning your help desk into a tool that AI assistants can securely work with directly. Point an AI agent at your HelpSpot MCP endpoint and it can search tickets, draft and update requests, browse and write knowledge base content, and pull reporting data — all on your behalf.

Connect the tools you already use

The server speaks standard MCP over HTTPS, so it works with popular AI clients out of the box:

• Claude: Claude Desktop and Claude Code
• ChatGPT: via the ChatGPT web app’s custom connector support (Developer Mode).
• Cursor and other MCP-compatible editors and agents.

Secure by design

• Off by default: an administrator must explicitly enable the MCP server before any agent can connect.
• OAuth or API tokens: connect with one-click OAuth or a personal API token generated from Preferences.
• Inherits your permissions: an agent can only see and do what the authenticating user can.
• Read-only mode: admins can hide all write tools (create/update) so agents can look but not touch.

What agents can do

The server exposes a full set of tools, grouped by area:

• Requests: get a single request or fetch many at once, view request history, full-text search, run your saved filters, and create or update both public and private requests (with clickable links back into HelpSpot).
• Knowledge base: list books and chapters, read pages, full-text search the KB, and create books, chapters, and pages or update existing pages.
• Reporting: aggregate requests and request history for activity and trend reporting, using the same date/status/category filters as search.
• Discovery: list categories (with reporting tags), statuses, staff, custom fields, and saved filters so agents can map your HelpSpot before acting.

Setup instructions for each client are in the bundled MCP Setup Guide.

Other New Features

• Set request status from the email parser: A new ##hs_status:N## parser tag lets incoming email set a request’s initial status, mirroring the existing ##hs_category:N## and ##hs_assigned_to:N## tags. A valid active status overrides the default; spam classification still takes precedence, so senders can’t use it to bypass the spam queue. See Email Parser: Advanced Tags for full documentation.

Bug Fixes

• Fixed email send failure on certain inline-image filenames: Inline attachments whose filenames had no real extension but contained dots/commas (e.g. addresses like “Acme Corp., La Vista, NE 68128”) produced a malformed Content-ID that violated RFC 2822, causing the entire email to fail on send. The extension is now sanitized to alphanumerics before being used in the Content-ID.