Provides tools for interacting with Zendesk, enabling ticket search and analysis, and Help Center article management (list, search, create, update).
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Zendesk MCP Server Fetch all tickets from the past week and analyze trends"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A Zendesk MCP (Model Context Protocol) server for Kojo. Gives Claude access to support tickets and Help Center articles for trend analysis, knowledge base management, and customer insight research.
"Fetch all tickets from the past week and provide an analysis of any trends you find"
"Pull last month's solved tickets and identify the top 5 feature gaps driving support volume"
"Search our help center for articles about approvals — are there any gaps we should fill?"
"Draft a new help center article explaining how to set up invoice matching"
Related MCP server: Zendesk MCP Server
Tool | Description |
| Search tickets by date range and status. Returns full conversation transcripts. Supports natural language dates ("past month", "past 7 days"). |
| Search, list, or get Help Center articles. Includes section discovery for browsing the knowledge base structure. |
| Create new Help Center articles (drafts by default). |
| Update article title, body, or draft status. |
Log into Zendesk as an admin
Go to Admin Center > Apps and integrations > Zendesk API
Enable Token Access if not already enabled
Click Add API token, give it a name, and copy the token
git clone <repo-url>
cd zendesk-mcp
npm install
npm run buildAdd to your .mcp.json:
{
"mcpServers": {
"zendesk": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/zendesk-mcp/build/index.js"],
"env": {
"ZENDESK_SUBDOMAIN": "your-subdomain",
"ZENDESK_EMAIL": "your-email@example.com",
"ZENDESK_API_TOKEN": "your-api-token"
}
}
}
}Replace the env values with your own. The subdomain is the part before .zendesk.com in your Zendesk URL.
Variable | Description |
| Default permission group for new articles. If unset, Zendesk assigns its own default. |
Node.js 18+
A Zendesk account with API token access
Claude Code with MCP support
Create a new article in the Zendesk Help Center. The article is created as a draft by default to allow review before publishing. Requires a section_id — use read_articles with include_sections=true to discover available sections.
| Name | Required | Description | Default |
|---|---|---|---|
| body | Yes | Article body content. Supports HTML formatting. | |
| draft | No | Whether the article should be created as a draft (default: true for safety). | |
| title | Yes | Article title. | |
| locale | No | Article locale (default: "en-us"). | en-us |
| section_id | Yes | ID of the section to create the article in. Use read_articles with include_sections=true to discover section IDs. | |
| label_names | No | Labels to apply to the article. | |
| notify_subscribers | No | Whether to notify section subscribers (default: false). | |
| permission_group_id | No | Permission group ID. If omitted, uses the default. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond the annotations (readOnly=false, idempotent=false), the description discloses a key behavioral trait: 'The article is created as a draft by default to allow review before publishing.' This adds context not captured in the annotation hints, though it doesn't detail error cases or response structure.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences, front-loaded with the core purpose, and contains no redundant words. Every sentence earns its place, with the second sentence providing crucial usage guidance.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For an 8-parameter tool with no output schema, the description covers the essential context: the action, the draft default, and the prerequisite section_id discovery. However, it does not mention what the response contains (e.g., the created article with ID), which would improve completeness. Still, the schema is rich and the guidance is solid.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
All 8 parameters are documented in the schema (100% coverage), so the baseline is 3. The description only restates the draft default and section_id requirement, which are already present in the schema descriptions. It adds no new parameter-level semantics.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states 'Create a new article in the Zendesk Help Center', identifying the specific action (create) and resource (article). It differentiates from siblings like update_article (modifies existing) and read_articles (reads), making the purpose unambiguous.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly says 'Requires a section_id — use read_articles with include_sections=true to discover available sections.' This tells the agent exactly when to use this tool and how to obtain a prerequisite using a sibling, which is strong guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Fetch Zendesk support tickets with full conversation transcripts for trend analysis. Filters by date range and ticket status. Returns ticket metadata (subject, status, priority, tags, dates) and the complete conversation transcript for each ticket. Use for analyzing support trends, common issues, and customer sentiment.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | Maximum number of tickets to return (default: 100, max: 1000). Higher limits take longer due to per-ticket comment fetching. | |
| status | No | Filter by ticket status. Provide one or more statuses. Example: ["open", "pending"]. If omitted, all statuses are returned. | |
| end_date | No | End date filter. Same formats as start_date. Only tickets created on or before this date are returned. Defaults to today if start_date is provided. | |
| start_date | No | Start date filter. Supports relative dates like "past month", "past week", "past 7 days", "today", "yesterday", or ISO dates like "2026-03-01". Only tickets created on or after this date are returned. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint and idempotentHint, and the description adds meaningful behavioral detail by explaining that the tool returns ticket metadata plus complete conversation transcripts, and supports date/status filters. There is no contradiction with the annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is three sentences, front-loaded with the main purpose, then output specifics, and a stated use case. Every sentence earns its place and there is no redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a read-only fetch tool with no output schema, the description explains the return content (metadata and transcripts), filters, and usage context. Combined with rich schema descriptions and annotations, there are no major contextual gaps.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with each parameter fully described, including defaults and accepted formats. The description only restates the filtering capability without adding additional parameter-level detail, so the schema carries the semantic burden.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb and resource ('Fetch Zendesk support tickets') and immediately distinguishes the tool by its key feature: full conversation transcripts. It clearly differentiates itself from the article-focused sibling tools by domain and purpose.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly states the intended use case: 'Use for analyzing support trends, common issues, and customer sentiment.' It doesn't mention when not to use it or name alternatives, but the context is clear and sufficient for an agent.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read articles from the Zendesk Help Center (Guide). Supports four modes: "search" for full-text search, "list" to browse articles by section or labels, "get" to retrieve a single article by ID with full content, and "sections" to list all Help Center sections (lightweight — use this to discover section IDs before creating articles).
| Name | Required | Description | Default |
|---|---|---|---|
| mode | Yes | Operation mode. "search": full-text search across articles. "list": browse articles, optionally filtered by section or labels. "get": retrieve a single article by ID with full content. "sections": list all Help Center sections (lightweight, no articles fetched). | |
| query | No | Search query string. Required when mode is "search". | |
| locale | No | Locale code (e.g., "en-us"). Defaults to the Help Center default locale. | |
| sort_by | No | Sort field. Used with mode "list". Default: position. | |
| per_page | No | Number of articles to return (max 100, default 25). | |
| article_id | No | Article ID. Required when mode is "get". | |
| section_id | No | Filter articles by section ID. Used with mode "list". | |
| sort_order | No | Sort order. Default: asc. | |
| label_names | No | Comma-separated label names to filter by. Used with "search" and "list" modes. | |
| include_sections | No | When true, also returns a list of all Help Center sections. Useful for discovering section IDs before creating articles. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already mark this as read-only and idempotent, so the safety profile is known. The description adds behavioral nuance by labeling 'sections' as 'lightweight' and explaining that 'get' retrieves 'full content,' which helps the agent understand performance and output characteristics without contradicting the annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single, information-dense sentence that front-loads the core function and uses a colon to enumerate the four modes. Each mode has a brief but meaningful explanation, and there is no redundant filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a read tool with strong annotations and a fully-described schema, the description covers the main usage scenarios and explicitly links to the article-creation workflow. It lacks details about return values, but the absence of an output schema and the read-only nature make this acceptable.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with detailed parameter descriptions. The tool description adds a mode-oriented overview (e.g., 'search' for full-text, 'list' to browse by section or labels) that helps map parameters to modes, but it doesn't provide additional syntax beyond what the schema already documents.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb and resource: 'Read articles from the Zendesk Help Center (Guide).' It then details four distinct modes, making it clear this is a read tool for articles, distinguishing it from sibling write tools like create_article and update_article.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit use cases for each mode, notably the sections mode: 'use this to discover section IDs before creating articles,' which directly ties to the sibling create_article tool. While it doesn't explicitly contrast with fetch_tickets, the article focus and write-tool siblings make the context clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Update an existing Zendesk Help Center article's content (title, body, draft status). Uses the Zendesk Translations API to modify content for the specified locale. At least one of title, body, or draft must be provided.
| Name | Required | Description | Default |
|---|---|---|---|
| body | No | New article body content (HTML). Omit to keep current body. | |
| draft | No | Set to true to make article a draft, false to publish. Omit to keep current status. | |
| title | No | New article title. Omit to keep current title. | |
| locale | No | Locale of the translation to update (default: "en-us"). | en-us |
| article_id | Yes | ID of the article to update. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already indicate it is not read-only and is idempotent, but the description adds useful behavioral context: it uses the Zendesk Translations API and operates on a specific locale. It also discloses a usage requirement (at least one field) that is not present in the annotations. It does not mention permissions or error handling, but the bar is lower due to existing annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is only two sentences long and front-loads the core purpose in the first sentence. The second sentence provides an essential usage constraint without any unnecessary words. Every phrase adds value, making it both concise and well-structured.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the moderate complexity (5 parameters, no output schema), the description covers purpose, the API used, and a key operational constraint. It does not specify the return value or potential errors, but the combination of annotations and schema covers most of the necessary context, making it suitably complete for an agent to select and invoke correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema describes all parameters, but the description adds a crucial collective requirement ('At least one of title, body, or draft must be provided') that is not obvious from the schema alone. It also clarifies the role of the locale parameter by mentioning the Translations API, enhancing the value beyond the individual parameter descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('Update') and the specific resource ('existing Zendesk Help Center article') along with the content fields affected (title, body, draft status). This distinguishes it from siblings like 'create_article' (create) and 'read_articles' (read), leaving no ambiguity about what the tool does.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides clear context that this tool is for modifying existing articles and includes a specific constraint ('At least one of title, body, or draft must be provided'). It does not explicitly name sibling alternatives, but the 'existing' wording and the tool's purpose imply when it should be used versus create/read tools, so it nearly meets the 5 threshold.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v1.0.0create_articlefetch_ticketsread_articlesupdate_articleScored across 4 tools
Each tool has a clearly distinct purpose: fetch_tickets is exclusively for support tickets, while read_articles, create_article, and update_article handle help center articles without overlap. Even though read_articles supports multiple modes, it does not conflict with the write tools.
All tool names follow a consistent verb_noun snake_case pattern (fetch_tickets, read_articles, create_article, update_article). The verbs are action-oriented and clearly indicate the operation, making the naming predictable and easy to navigate.
With only 4 tools, the set is compact but well-scoped for the two main domains (ticket analysis and article management). Each tool serves a necessary function without redundancy or bloat.
The article workflow is nearly complete, with read, create, and update operations, but it lacks delete functionality. Ticket handling only supports fetching, which aligns with the stated trend-analysis purpose but may be limiting for other use cases. Overall, the coverage is solid with minor gaps.
Read tickets, users, orgs, macros and satisfaction ratings; create, update and comment on tickets.
Read tickets, contacts, companies, agents and groups; create, update and reply to tickets.
CX ops: read conversations, calls and QA evaluations from Zendesk, Freshdesk, Five9 and more.
Zendesk MCP Pack — tickets, users, organizations via OAuth.
