AgenticBI
Connect your business data to any MCP client. AgenticBI agents find the right data, join across systems, run queries, build dashboards, generate reports, and automate analytics workflows. No SQL required.
README
AgenticBI MCP Server
Demo
Watch AgenticBI in action: https://www.youtube.com/@TryAgenticBI
| Field | Value |
|---|---|
| Product | AgenticBI MCP Server |
| Purpose | Connect Claude, Cursor, VS Code, ChatGPT, Codex CLI, and other MCP-enabled AI clients directly to business data |
| Endpoint | https://app.agenticbi.com/api/2.0/mcp |
| Protocol | MCP 2024-11-05 |
| Transport | Streamable HTTP (POST) with SSE |
| Authentication | OAuth 2.1 with PKCE |
| Tools Available | 33 |
| GitHub | https://github.com/AgenticBIHQ/agenticbi-mcp-server |
Prerequisites
| Requirement |
|---|
| MCP-compatible AI client (Claude Desktop, Claude Code, Cursor, VS Code, ChatGPT, Codex CLI) |
Client Setup
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"agenticbi": {
"url": "https://app.agenticbi.com/api/2.0/mcp"
}
}
}
File locations:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
On first use, your browser opens to AgenticBI login. Sign in or create an account to authorize.
Full Guide: https://app.agenticbi.com/docs/mcp-server.html
Claude Code
claude mcp add agenticbi --transport http https://app.agenticbi.com/api/2.0/mcp
-
Run the command above in your terminal.
-
Exit and relaunch Claude Code.
-
Run
/mcp, select agenticbi, and click Enable. -
Click the authentication button — your browser opens to AgenticBI login.
-
Sign in or create an account and authorize access.
-
Verify by asking:
What AgenticBI tools are available?
Alternative: Manual Token Authentication
Generate a token in AgenticBI → Settings → AI Settings → MCP Token, then run:
claude mcp add agenticbi https://app.agenticbi.com/api/2.0/mcp \
--transport http \
--header "Authorization: Bearer YOUR_MCP_TOKEN"
Full Guide: https://app.agenticbi.com/docs/claude-code-setup.html
Cursor
Requires Cursor v0.45 or later.
- Open Settings → Cursor Settings
- Select Tools & Integrations → New MCP Server
- Add the following to your
mcp.json:
{
"mcpServers": {
"agenticbi": {
"url": "https://app.agenticbi.com/api/2.0/mcp"
}
}
}
- Save the file. Your browser opens to AgenticBI login on first use.
Alternative: Manual Token Authentication
{
"mcpServers": {
"agenticbi": {
"url": "https://app.agenticbi.com/api/2.0/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}
Full Guide: https://app.agenticbi.com/docs/cursor-mcp-setup.html
VS Code
Option A — Command Palette (Recommended)
-
Open Command Palette:
- macOS:
Cmd + Shift + P - Windows:
Ctrl + Shift + P
- macOS:
-
Search for MCP: Add Server
-
Select HTTP as the server type
-
Enter:
https://app.agenticbi.com/api/2.0/mcp -
Name it
agenticbi -
Save
-
Authorize via browser on first use
Option B — Configuration File
Add to settings.json:
{
"mcp": {
"servers": {
"agenticbi": {
"type": "http",
"url": "https://app.agenticbi.com/api/2.0/mcp"
}
}
}
}
Or add to .vscode/mcp.json:
{
"servers": {
"agenticbi": {
"type": "http",
"url": "https://app.agenticbi.com/api/2.0/mcp"
}
}
}
Alternative: Manual Token Authentication
{
"mcp": {
"servers": {
"agenticbi": {
"type": "http",
"url": "https://app.agenticbi.com/api/2.0/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}
}
Full Guide: https://app.agenticbi.com/docs/vscode-mcp-setup.html
ChatGPT
-
Click + in the message input and select Connectors, or go to Settings → Connected Apps
-
Choose Add Custom Connector
-
Name it AgenticBI
-
Enter the server URL:
https://app.agenticbi.com/api/2.0/mcp -
Select OAuth 2.0 authentication
-
Click Save
-
On first use, ChatGPT opens your browser for AgenticBI sign-in and authorization
Alternative: Manual Token Authentication
- Generate a token in AgenticBI → Settings → AI Settings → MCP Token
- Repeat steps 1–4 above
- Select Bearer Token authentication
- Paste your token
- Click Save
Full Guide: https://app.agenticbi.com/docs/chatgpt-mcp-setup.html
Codex CLI
Add to ~/.codex/config.json:
{
"mcpServers": {
"agenticbi": {
"type": "http",
"url": "https://app.agenticbi.com/api/2.0/mcp"
}
}
}
On first use, your browser opens to AgenticBI login for authorization.
Alternative: Manual Token Authentication
{
"mcpServers": {
"agenticbi": {
"type": "http",
"url": "https://app.agenticbi.com/api/2.0/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}
Full Guide: https://app.agenticbi.com/docs/codex-mcp-setup.html
Authentication Endpoints
| Endpoint | URL |
|---|---|
| Authorization | https://app.agenticbi.com/oauth/mcp/authorize |
| Token | https://app.agenticbi.com/oauth/mcp/token |
| Registration | https://app.agenticbi.com/oauth/mcp/register |
Tool Categories
| Category | Tool Count |
|---|---|
| Ask & Query | 6 |
| Dashboards & Exports | 4 |
| Insights & Alerts | 3 |
| Documents | 2 |
| Create & Manage | 16 |
| Destructive | 2 |
| Total | 33 |
Ask & Query Tools
| Tool | Description |
|---|---|
ask |
Ask a data question in natural language |
get_data |
Retrieve full dataset or widget data |
search |
Search AgenticBI assets |
find_widget |
Find widgets by name |
get_details |
Retrieve asset details |
explain_lineage |
Trace how an answer was built |
Dashboard & Export Tools
| Tool | Description |
|---|---|
list_dashboards |
List dashboards |
get_embed_url |
Generate embed links |
export_pdf |
Export dashboards/widgets as PDF |
export_csv |
Export data as CSV |
Insights & Alerts Tools
| Tool | Description |
|---|---|
get_insights |
Generate AI-powered insights |
list_alerts |
List alerts |
cloud9ql_assist |
Generate, fix, explain, convert, validate or preview Cloud9QL |
Document Tools
| Tool | Description |
|---|---|
ask_document |
Ask questions against ingested documents |
list_documents |
List ingested documents |
Create & Manage Tools
| Tool | Description |
|---|---|
do |
Execute operations in natural language |
push_data |
Push rows into datasets |
run |
Execute saved queries/reports |
create_datasource |
Create a datasource connection |
create_query |
Create a query |
create_widget |
Create a visualization |
update_widget |
Update a visualization |
layout_dashboard |
Arrange dashboard widgets |
import_file |
Import CSV, JSON, XML, or Excel |
create_report |
Create scheduled reports |
update_report |
Update reports |
create_alert |
Create alerts |
test_alert |
Test alerts |
update_alert |
Update alerts |
ingest_document |
Ingest documents |
extract_document_data |
Extract structured data from documents |
Destructive Tools
| Tool | Description |
|---|---|
delete |
Delete AgenticBI assets |
delete_document |
Delete ingested documents |
Health & Limits
| Setting | Value |
|---|---|
| Health Endpoint | GET /api/2.0/mcp/health |
| Session Header | Mcp-Session-Id |
| Idle Timeout | 30 minutes |
| Max Concurrent Sessions | 1,000 |
Default Rows (ask) |
500 |
Default Rows (get_data) |
10,000 |
| Max Rows | 200,000 |
Example Prompts
| Example |
|---|
| What were our top 10 customers by revenue last quarter? |
| Why did churn spike in March? |
| Which marketing channels drove the most pipeline this month? |
| Create a dashboard showing weekly signups by source. |
| Alert me when MRR drops more than 5% week over week. |
| Pull our call transcripts and summarize the top objections from lost deals. |
| Upload this PDF and extract the key metrics into a dataset. |
| Export the Q3 revenue dashboard as a PDF and email it to the team. |
Important Links
| Resource | URL |
|---|---|
| Website | https://www.agenticbi.com |
| MCP Docs | https://app.agenticbi.com/docs/mcp-server.html |
| Privacy Policy | https://www.agenticbi.com/privacy |
| Terms of Service | https://www.agenticbi.com/terms |
Support
| Contact |
|---|
| support@agenticbi.com |
Directory Listings
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。