rush-db · 1pxone · May 14, 2026 · May 17, 2026
diff --git a/.changeset/strong-buses-burn.md b/.changeset/strong-buses-burn.md
@@ -0,0 +1,10 @@
+---
+'rushdb-dashboard': patch
+'rushdb-docs': patch
+'@rushdb/javascript-sdk': patch
+'@rushdb/mcp-server': patch
+'@rushdb/skills': patch
+'rushdb-core': patch
+---
+
+Dashboard help panel improvements
diff --git a/docs/docs/mcp-server/configuration.mdx b/docs/docs/mcp-server/configuration.mdx
@@ -7,16 +7,32 @@ sidebar_position: 4
 
 # Configuration
 
-| Variable | Required | Default | Description |
-|---|---|---|---|
-| `RUSHDB_API_KEY` | **yes** | — | Your RushDB API token (from the dashboard) |
-| `RUSHDB_API_URL` | no | `https://api.rushdb.com/api/v1` | Override for self-hosted or staging |
+## Connection modes
+
+### Remote OAuth (ChatGPT, Claude.ai, web clients)
+
+No configuration file needed. Point your AI client at:
+
+```
+https://mcp.rushdb.com/mcp
+```
+
+Authentication is handled via OAuth — your client will redirect to RushDB for sign-in. No API key is required in this mode; you authorize with your RushDB account and the server issues a token scoped to the selected project.
+
+### Local (API key)
+
+For stdio-based clients (Claude Desktop, Cursor, VS Code), the MCP package is configured through environment variables:
+
+| Variable         | Required | Default                         | Description                                |
+| ---------------- | -------- | ------------------------------- | ------------------------------------------ |
+| `RUSHDB_API_KEY` | **yes**  | —                               | Your RushDB API token (from the dashboard) |
+| `RUSHDB_API_URL` | no       | `https://api.rushdb.com/api/v1` | Override for self-hosted or staging        |
 
 :::tip Security
 Treat `RUSHDB_API_KEY` like a password — do not commit it to version control. Prefer MCP client `env` injection or OS keychains over `.env` files in repositories.
 :::
 
-## Providing values
+## Providing API key values
 
 ### 1) MCP client configuration
 

diff --git a/docs/docs/mcp-server/introduction.mdx b/docs/docs/mcp-server/introduction.mdx
@@ -7,7 +7,7 @@ sidebar_position: 1
 
 # MCP Server
 
-Connect Claude, Cursor, or any MCP-compatible client to your RushDB database. Type natural language; the server handles discovery, query building, and execution.
+Connect Claude, ChatGPT, Cursor, or any MCP-compatible client to your RushDB database. Type natural language; the server handles discovery, query building, and execution.
 
 ```
 You: How many MOVIE records have a rating above 8?
@@ -18,6 +18,18 @@ Claude: [calls getOntologyMarkdown → confirms MOVIE label and rating field]
 Result: 1 movie matched (Inception, rating 8.8)
 ```
 
+## Connection modes
+
+**Remote OAuth** — connect directly from ChatGPT, Claude.ai, or any web-based AI assistant. No local installation required. Authenticates via your RushDB account at:
+
+```
+https://mcp.rushdb.com/mcp
+```
+
+**Local (API key)** — install `@rushdb/mcp-server` in Claude Desktop, Cursor, VS Code, or any stdio-based MCP client. Provide your RushDB API key as an environment variable.
+
+→ [Quickstart](./quickstart) for setup instructions for both modes.
+
 ## Mandatory workflow
 
 The MCP server ships with a built-in system prompt that enforces a four-step workflow:

diff --git a/docs/docs/mcp-server/quickstart.mdx b/docs/docs/mcp-server/quickstart.mdx
@@ -7,6 +7,39 @@ sidebar_position: 2
 
 # Quickstart
 
+Two ways to connect:
+
+- **Remote OAuth** — connect directly from ChatGPT, Claude.ai, or any web-based AI client with no local installation. Uses `https://mcp.rushdb.com/mcp` and authenticates via your RushDB account.
+- **Local (API key)** — install the MCP package in Claude Desktop, Cursor, VS Code, or any stdio-based client. Requires a RushDB API key.
+
+---
+
+## ChatGPT
+
+Connect RushDB to ChatGPT as a connector — no local installation required:
+
+1. In ChatGPT, open **Settings → Connectors → Add connector**
+2. Enter the MCP URL: `https://mcp.rushdb.com/mcp`
+3. Complete the OAuth sign-in with your RushDB account
+4. Select the project you want to expose to ChatGPT
+
+After linking, ChatGPT can query and write to your RushDB project directly.
+
+:::tip Troubleshooting
+If you see `OAuth linked, action discovery failed`, see the [Troubleshooting](./troubleshooting) page.
+:::
+
+---
+
+## Claude.ai
+
+1. In Claude.ai, open **Settings → Integrations → Add integration**
+2. Enter the MCP URL: `https://mcp.rushdb.com/mcp`
+3. Authorize with your RushDB account via OAuth
+4. The integration is now active across Claude.ai conversations
+
+---
+
 ## Claude Desktop
 
 Add to `~/Library/Application Support/Claude/claude_desktop_config.json`:
@@ -66,17 +99,22 @@ Ask Claude:
 
 **Override API URL** (self-hosted / staging): add `"RUSHDB_API_URL": "https://your-host/api/v1"` to any `env` block above.
 
-## Hosted MCP endpoint (connected clients)
+## Remote MCP endpoint
+
+The hosted endpoint for all remote and web-based clients:
 
-For connected clients that use an HTTP MCP endpoint, use:
+```
+https://mcp.rushdb.com/mcp
+```
 
-`https://mcp.rushdb.com/mcp`
+Do not use the root domain (`https://mcp.rushdb.com`) — always include `/mcp`.
 
-Do not use the root domain (`https://mcp.rushdb.com`) for connector setup.
+Local tunnel testing: `https://<your-ngrok-host>.ngrok-free.app/mcp`
 
-If you are testing locally through a tunnel, use the full tunnel endpoint path as well:
+OAuth discovery metadata is available at:
 
-`https://<your-ngrok-host>.ngrok-free.app/mcp`
+- `https://mcp.rushdb.com/.well-known/openid-configuration`
+- `https://mcp.rushdb.com/.well-known/jwks.json`
 
 ## Enable the Query Builder prompt (recommended)
 

diff --git a/docs/docs/tutorials/third-party-webhook-ingestion.mdx b/docs/docs/tutorials/third-party-webhook-ingestion.mdx
@@ -0,0 +1,198 @@
+---
+sidebar_position: 30
+title: 'Webhook Ingestion from Third-Party Tools (Clay, Supabase, Superglue)'
+description: 'A beginner-friendly guide to send JSON or CSV from no-code and data platforms into RushDB using custom HTTP requests.'
+tags: [Ingestion, Webhooks, Getting Started, No-Code]
+---
+
+# Webhook Ingestion from Third-Party Tools
+
+Many tools can send data to a custom HTTP endpoint. That means you can connect your existing workflow to RushDB without writing a full app.
+
+This guide is for non-technical teams and covers:
+
+- Clay
+- Supabase
+- Superglue
+- Any platform that supports custom HTTP requests or webhooks
+
+## What you need
+
+Before you start, prepare these values:
+
+- Your RushDB API key
+- Your RushDB API base URL: `https://api.rushdb.com`
+- A destination label (example: `LEAD`, `CUSTOMER`, `ORDER`)
+- A sample JSON or CSV payload
+
+You can get your API key from your dashboard project settings.
+
+## The universal webhook recipe
+
+Most platforms ask for the same settings. Use these defaults:
+
+- Method: `POST`
+- URL (JSON import): `https://api.rushdb.com/api/v1/records/import/json`
+- URL (CSV import): `https://api.rushdb.com/api/v1/records/import/csv`
+- Header: `Authorization: Bearer YOUR_RUSHDB_API_KEY`
+- Header: `Content-Type: application/json` (for JSON)
+- Body: include `label` and your data
+
+### JSON request body shape
+
+```json
+{
+  "label": "LEAD",
+  "data": [
+    {
+      "externalId": "lead_001",
+      "name": "Ava Collins",
+      "email": "ava@example.com",
+      "source": "Clay"
+    }
+  ]
+}
+```
+
+### CSV request body shape
+
+For CSV import, send your CSV text as request body and include a label parameter based on your platform's request builder. If your platform cannot send raw CSV, convert CSV rows to JSON and use the JSON endpoint.
+
+## Setup in Clay
+
+1. Open your table/workflow where rows are ready to export.
+2. Add an action step that can send an HTTP request (sometimes called Webhook or API request).
+3. Set method to `POST`.
+4. Set URL to `https://api.rushdb.com/api/v1/records/import/json`.
+5. Add headers:
+
+- `Authorization: Bearer YOUR_RUSHDB_API_KEY`
+- `Content-Type: application/json`
+
+6. Build the request body using Clay row fields and this structure:
+
+```json
+{
+  "label": "LEAD",
+  "data": [
+    {
+      "externalId": "{{row.id}}",
+      "name": "{{row.name}}",
+      "email": "{{row.email}}",
+      "company": "{{row.company}}"
+    }
+  ]
+}
+```
+
+7. Run a test send, then verify the new records in your RushDB project.
+
+## Setup in Supabase
+
+Supabase usually sends outbound HTTP from Edge Functions, triggers, or automation flows.
+
+1. Choose the event you want to sync (insert/update).
+2. Configure an outbound `POST` request to RushDB.
+3. Use URL: `https://api.rushdb.com/api/v1/records/import/json`.
+4. Add headers:
+
+- `Authorization: Bearer YOUR_RUSHDB_API_KEY`
+- `Content-Type: application/json`
+
+5. Map changed row fields into `data`.
+
+Optional payload example:
+
+```json
+{
+  "label": "CUSTOMER",
+  "data": [
+    {
+      "externalId": "{{new.id}}",
+      "email": "{{new.email}}",
+      "plan": "{{new.plan}}",
+      "updatedAt": "{{new.updated_at}}"
+    }
+  ]
+}
+```
+
+If your Supabase workflow exports CSV, route that payload to `https://api.rushdb.com/api/v1/records/import/csv` instead.
+
+## Setup in Superglue
+
+1. In your Superglue pipeline, add a destination step using custom HTTP.
+2. Select `POST` method.
+3. Choose JSON endpoint (`/api/v1/records/import/json`) for mapped objects or CSV endpoint (`/api/v1/records/import/csv`) for file/text streams.
+4. Add Authorization header with your RushDB API key.
+5. Map source fields to your target object shape.
+6. Send a test batch and confirm records appear in RushDB.
+
+Tip: start with a single small batch, validate fields, then scale.
+
+## JSON and CSV quick reference
+
+### Use JSON when
+
+- Your source already provides structured objects
+- You want field-level mapping control
+- You want easier debugging of each record
+
+### Use CSV when
+
+- Your source exports files or flat tables
+- You do not need nested object structures
+- You want fast bulk upload from tabular datasets
+
+## Troubleshooting
+
+### 401 or 403 errors
+
+- Check API key value and spacing in `Authorization` header
+- Ensure the key belongs to the correct project
+
+### 400 validation errors
+
+- Confirm `label` exists and is non-empty
+- Validate JSON syntax (trailing commas are common issues)
+- Check required fields and data types
+
+### Empty or malformed CSV imports
+
+- Ensure header row exists
+- Confirm consistent delimiter and quoting
+- Verify UTF-8 encoding
+
+### Duplicate records after retries
+
+Webhook systems may retry requests. Use an external stable ID (`externalId`) in each record so you can deduplicate and upsert safely.
+
+For stronger idempotency patterns, see [Event-Driven Ingestion from Webhooks and Queues](./event-driven-ingestion).
+
+## Optional: test from terminal first
+
+If you want to verify credentials before configuring a platform:
+
+```bash
+export RUSHDB_API_KEY="YOUR_RUSHDB_API_KEY"
+
+curl -X POST "https://api.rushdb.com/api/v1/records/import/json" \
+  -H "Authorization: Bearer $RUSHDB_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "label": "LEAD",
+    "data": [
+      {
+        "externalId": "lead_001",
+        "name": "Ava Collins",
+        "email": "ava@example.com"
+      }
+    ]
+  }'
+```
+
+## Related docs
+
+- [Importing data from external sources](./importing-data)
+- [Event-Driven Ingestion from Webhooks and Queues](./event-driven-ingestion)
+- [REST API introduction](../rest-api/introduction)