Qonto MCP Server
Enables interaction with the Qonto API v2 via 21 tools for organization, transactions, attachments, beneficiaries, clients, invoices, labels, memberships, requests, statements, and transfers, with GitHub OAuth authentication and stateful sessions.
README
Qonto MCP Server
A remote Model Context Protocol server for the Qonto API, deployed on Cloudflare Workers with GitHub OAuth authentication.
Features
- 21 tools covering the full Qonto API v2: organization, transactions, attachments, beneficiaries, clients, invoices, labels, memberships, requests, statements, and transfers
- OAuth 2.1 authentication via GitHub (Authorization Code + PKCE)
- Stateful sessions via Cloudflare Durable Objects
- Zero cold-start secrets — Qonto credentials never leave Cloudflare
Tools
| Category | Tools |
|---|---|
| Organization | get_qonto_organization |
| Transactions | get_qonto_transactions, get_qonto_transaction, list_qonto_transaction_attachments |
| Attachments | get_qonto_attachment |
| Beneficiaries | list_qonto_beneficiaries, get_qonto_beneficiary |
| Clients | get_clients, get_client |
| Invoices | get_client_invoices, get_supplier_invoices, get_credit_notes |
| Labels | list_qonto_labels, get_qonto_label |
| Memberships | list_qonto_memberships |
| Requests | get_requests, get_request |
| Statements | get_statements, download_statement |
| Transfers | list_qonto_external_transfers, get_qonto_external_transfer |
Prerequisites
- Node.js 18+
- Wrangler CLI (
npm install -g wrangler) - A Cloudflare account
- A Qonto account with API access (
Settings → Integrations → API) - A GitHub OAuth App
Setup
1. Clone and install
git clone git@github.com:thomas-rx/qonto-mcp-worker.git
cd qonto-mcp-worker
npm install
2. Create a GitHub OAuth App
Go to github.com/settings/developers → OAuth Apps → New OAuth App.
| Field | Value |
|---|---|
| Homepage URL | https://<worker-name>.<account>.workers.dev |
| Authorization callback URL | https://<worker-name>.<account>.workers.dev/callback |
For local development, create a second app with http://localhost:8788 and http://localhost:8788/callback.
3. Create the KV namespace
npx wrangler kv namespace create "OAUTH_KV"
Copy the returned ID into wrangler.jsonc:
"kv_namespaces": [{ "binding": "OAUTH_KV", "id": "<your-id>" }]
4. Set secrets
npx wrangler secret put QONTO_API_KEY
npx wrangler secret put QONTO_ORGANIZATION_ID
npx wrangler secret put GITHUB_CLIENT_ID
npx wrangler secret put GITHUB_CLIENT_SECRET
openssl rand -hex 32 | npx wrangler secret put COOKIE_ENCRYPTION_KEY
5. Local development
Copy .dev.vars.example to .dev.vars and fill in your credentials:
cp .dev.vars.example .dev.vars
npm run dev
# → http://localhost:8788
6. Deploy
npm run deploy
Connecting to Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"qonto": {
"command": "npx",
"args": [
"mcp-remote",
"https://<worker-name>.<account>.workers.dev/<mcp-path>"
]
}
}
}
On first connection, a browser window will open for GitHub authentication. Sessions persist for 30 days.
Architecture
Request → OAuthProvider
├── /authorize → GitHub OAuth consent flow
├── /callback → Token exchange + Qonto session
├── /token → OAuth token endpoint
├── /register → Dynamic client registration
└── /<mcp-path> → QontoMCP (Durable Object)
└── QontoClient → Qonto API v2
Environment Variables
| Variable | Required | Description |
|---|---|---|
QONTO_API_KEY |
Yes | Qonto secret key |
QONTO_ORGANIZATION_ID |
Yes | Qonto organization slug |
QONTO_THIRDPARTY_HOST |
No | Override API base URL (default: https://thirdparty.qonto.com) |
QONTO_STAGING_TOKEN |
No | Qonto staging token |
GITHUB_CLIENT_ID |
Yes | GitHub OAuth App client ID |
GITHUB_CLIENT_SECRET |
Yes | GitHub OAuth App client secret |
COOKIE_ENCRYPTION_KEY |
Yes | 32-byte hex key for signing cookies |
License
MIT
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。