Productboard MCP Server
MCP server for integrating with Productboard to manage customer feedback, fetch notes, search and create features, and link notes to features.
README
Productboard MCP Server
MCP server that lets Claude Code fetch notes, search features, create features, and link notes to features in Productboard. Includes a triage skill for processing customer feedback.
Setup
- Clone and build:
git clone git@github.com:djcorbett96/kojo-productboard-mcp.git
cd kojo-productboard-mcp
npm install
npm run build
-
Get a Productboard API token from your workspace admin settings under Integrations > Notes API.
-
Add to your Claude Code MCP config (
.mcp.jsonin your project root or~/.claude/mcp.jsonfor global):
{
"mcpServers": {
"productboard": {
"command": "node",
"args": ["/path/to/kojo-productboard-mcp/build/index.js"],
"env": {
"PRODUCTBOARD_API_TOKEN": "your-token-here",
"PRODUCTBOARD_TRIAGE_COMPONENT_ID": "",
"PRODUCTBOARD_DEFAULT_STATUS_ID": ""
}
}
}
}
| Env Variable | Required | Description |
|---|---|---|
PRODUCTBOARD_API_TOKEN |
Yes | Your Productboard API token |
PRODUCTBOARD_TRIAGE_COMPONENT_ID |
For feature creation | Component ID where new triage features are created. Find this in Productboard's feature board URL or via the API. |
PRODUCTBOARD_DEFAULT_STATUS_ID |
For feature creation | Status ID for new features (e.g., "New idea"). Use GET /feature-statuses with X-Version: 1 header to list available statuses. |
Claude Code launches the server automatically -- no need to run it manually.
Tools
fetch_notes
Fetches notes from Productboard with optional filtering.
| Parameter | Type | Description |
|---|---|---|
tags |
string[] | Filter to notes containing at least one of these tags |
keywords |
string | Full-text search on note titles and content |
createdFrom |
string | Notes created on/after this date -- "past month", "past week", "past 7 days", or "2024-01-15" |
state |
string | Filter by processing state: "processed" or "unprocessed" |
limit |
number | Max notes to return (default: 100, max: 1000) |
All parameters are optional. The server handles pagination automatically.
search_features
Searches existing features by keyword. Paginates through all features and performs client-side matching on names and descriptions.
| Parameter | Type | Description |
|---|---|---|
query |
string | Search term to match against feature names and descriptions (required) |
limit |
number | Max matching features to return (default: 25, max: 100) |
create_feature
Creates a new feature in Productboard. Requires PRODUCTBOARD_TRIAGE_COMPONENT_ID and PRODUCTBOARD_DEFAULT_STATUS_ID env vars.
| Parameter | Type | Description |
|---|---|---|
name |
string | Feature name (required) |
description |
string | Feature description (supports HTML) |
parentComponentId |
string | Override the default triage component |
statusId |
string | Override the default status |
link_note_to_feature
Links a customer feedback note to a feature for tracking and prioritization.
| Parameter | Type | Description |
|---|---|---|
noteId |
string | The note ID to link (required) |
featureId |
string | The feature ID to link to (required) |
Triage Skill
The repo includes a Claude Code skill for triaging unprocessed Productboard notes. It fetches notes, matches them to existing features, proposes new features when needed, and waits for your approval before taking action.
Skill Setup
Copy the skill into your project's Claude Code skills directory:
cp -r /path/to/kojo-productboard-mcp/.claude/skills/triage /your/project/.claude/skills/triage
Or symlink it:
ln -s /path/to/kojo-productboard-mcp/.claude/skills/triage /your/project/.claude/skills/triage
Usage
Run the skill in Claude Code with your product area tags:
/triage Warehouse, Procurement
/triage AP past month
/triage Platform, Integrations past 2 weeks
The skill will:
- Fetch unprocessed notes matching your tags
- Search for existing features that match each note
- Present a triage report with proposed actions (link to existing feature, create new feature, or flag for manual review)
- Wait for your approval before creating any features or linking any notes
The fetch_notes and search_features calls are pre-approved in the skill config. You'll only be prompted to approve create_feature and link_note_to_feature calls.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。