Productive.io MCP Server
Enables AI assistants to interact with a Productive.io workspace for managing projects, tasks, time entries, budgets, and invoices through natural language.
README
Productive.io MCP Server
An MCP (Model Context Protocol) server for Productive.io that enables AI assistants like Claude to interact with your Productive.io workspace — managing projects, tasks, time entries, budgets, and more.
Supports two transport modes:
- stdio (default) — Claude launches it as a local subprocess
- HTTP (with
--httpflag) — runs as an Express web server for remote/team use
Prerequisites
- Node.js 18+
- A Productive.io account with API access
- API Token — generate one in Productive.io under Settings → API integrations
- Organization ID — found in your Productive.io account settings
Installation
git clone <your-repo-url> productive-mcp-server
cd productive-mcp-server
npm install
cp .env.example .env
# Edit .env with your credentials
npm run build
Configuration
| Variable | Required | Description |
|---|---|---|
PRODUCTIVE_API_TOKEN |
Yes | Your Productive.io API token |
PRODUCTIVE_ORG_ID |
Yes | Your Productive.io organization ID |
PORT |
No | HTTP server port (default: 3000) |
Running
stdio mode (default — for local use)
# Development
npm run dev
# Production (after build)
npm start
HTTP mode (for remote/team use)
# Development
npm run dev:http
# Production (after build)
npm run start:http
The HTTP server exposes:
POST /mcp— MCP protocol endpointGET /health— health check
Connecting to Claude Desktop
Add this to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"productive": {
"command": "node",
"args": ["/absolute/path/to/productive-mcp-server/dist/index.js"],
"env": {
"PRODUCTIVE_API_TOKEN": "your_api_token_here",
"PRODUCTIVE_ORG_ID": "your_org_id_here"
}
}
}
}
Connecting to Claude Code
stdio mode
claude mcp add productive -- node /absolute/path/to/productive-mcp-server/dist/index.js
Set the environment variables before running Claude Code, or add them to your shell profile:
export PRODUCTIVE_API_TOKEN=your_api_token_here
export PRODUCTIVE_ORG_ID=your_org_id_here
HTTP mode
Start the server first:
cd /path/to/productive-mcp-server
PRODUCTIVE_API_TOKEN=your_token PRODUCTIVE_ORG_ID=your_org npm run start:http
Then connect Claude Code:
claude mcp add productive --transport http http://localhost:3000/mcp
Connecting to VS Code
Add to your VS Code user settings JSON (.vscode/settings.json or user settings):
stdio mode
{
"mcp": {
"servers": {
"productive": {
"command": "node",
"args": ["/absolute/path/to/productive-mcp-server/dist/index.js"],
"env": {
"PRODUCTIVE_API_TOKEN": "your_api_token_here",
"PRODUCTIVE_ORG_ID": "your_org_id_here"
}
}
}
}
}
HTTP mode
{
"mcp": {
"servers": {
"productive": {
"type": "http",
"url": "http://localhost:3000/mcp"
}
}
}
}
Available Tools
| Tool | Description | Key Parameters |
|---|---|---|
productive_list_projects |
List/search projects | company_id, project_type_id, status, query |
productive_get_project |
Get project by ID | id, include |
productive_create_project |
Create a project | name, project_type_id, company_id |
productive_update_project |
Update a project | id, name, project_type_id |
productive_list_tasks |
List/filter tasks | project_id, assignee_id, status, query |
productive_get_task |
Get task by ID | id, include |
productive_create_task |
Create a task | title, task_list_id, assignee_id |
productive_update_task |
Update a task | id, title, closed, assignee_id |
productive_list_time_entries |
List time entries | person_id, project_id, after, before |
productive_log_time |
Log a time entry | person_id, service_id, date, time |
productive_list_people |
List team members | status, company_id, query |
productive_get_person |
Get person by ID | id, include |
productive_list_deals |
List budgets/deals | project_id, company_id, deal_status |
productive_get_deal |
Get deal by ID | id, include |
productive_list_companies |
List companies | status, query |
productive_get_company |
Get company by ID | id, include |
productive_list_bookings |
List bookings | person_id, project_id, after, before |
productive_list_services |
List services | deal_id |
Invoicing Tools (Draft Only)
All invoice tools create and manage draft invoices only. Finalization (assigning an invoice number) must always be done manually in the Productive.io UI.
| Tool | Description | Key Parameters |
|---|---|---|
productive_list_invoices |
List/filter invoices | company_id, deal_id, invoice_status, date range |
productive_get_invoice |
Get full invoice details with line items | id |
productive_create_invoice |
Create a new empty draft invoice | company_id, document_type_id, subsidiary_id |
productive_update_invoice |
Update a draft invoice | id, subject, note, dates |
productive_create_invoice_from_previous |
Clone an invoice (line items + budget links) | source_invoice_id, optional overrides |
productive_create_invoice_like_last_for_client |
Clone a client's most recent invoice | company_id, how_many_back |
productive_list_line_items |
List line items for an invoice | invoice_id |
productive_create_line_item |
Add a line item to a draft invoice | invoice_id, description, quantity, unit_price, unit_id |
productive_update_line_item |
Update an existing line item | id, description, quantity, unit_price |
productive_delete_line_item |
Delete a line item | id |
productive_list_invoice_attributions |
List budget links for an invoice | invoice_id |
productive_create_invoice_attribution |
Link a draft invoice to a budget | invoice_id, deal_id, amount |
productive_list_document_types |
List available document types | subsidiary_id, status |
productive_list_subsidiaries |
List your company subsidiaries | status |
productive_list_tax_rates |
List available tax rates | status |
All list tools support page, page_size, and sort parameters.
Testing with MCP Inspector
npx @modelcontextprotocol/inspector node dist/index.js
This opens an interactive UI to test each tool.
Rate Limits
Productive.io enforces:
- 100 requests per 10 seconds
- 4,000 requests per 30 minutes
The server implements automatic retry with exponential backoff for 429 responses (up to 3 retries).
License
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。