Model Context Protocol (JSON-RPC 2.0) Active

Machine-to-Machine Protocols

Equip Claude Desktop, Cursor IDE, LangChain, or external AI agent swarms directly with 16 live business tools and autonomous sub-agent task routing via Model Context Protocol (MCP) and A2A.

How MCP Integration Works

JSON-RPC 2.0 protocol cycle between external AI host and Bless Web Designs tools.

STEP 1 CLIENT

🔌 Claude / Cursor / Agent

External AI host connects via HTTP POST and issues a JSON-RPC 2.0 request with x-api-key.

POST /mcp
STEP 2 HANDSHAKE

📜 Discovery & Scope

Client calls tools/list to dynamically discover tool schemas (or calls tools/call).

method: "tools/list"
STEP 3 ROUTING

🔀 Tool or Agent Loop

Direct tools run handlers immediately. If agent_message is called, it triggers a full sub-agent reasoning loop.

agent_message & 16 Tools
STEP 4 EXECUTION

⚙️ Live Business Actions

Executes on-page SEO crawling, Calendly availability, ClickUp tasks, or QuickBooks invoices.

Safe Sandboxed Handlers
STEP 5 RETURN

📦 JSON-RPC Result

Returns standardized JSON-RPC 2.0 result containing tool output or synthesized sub-agent reply.

{ result: { ... } }

Live MCP Protocol Testbed

Send real JSON-RPC 2.0 calls directly to POST /mcp and inspect responses in real time.

🔌 MCP JSON-RPC 2.0 Dispatcher • Server Port 8787
Gateway Active
Ready
🔌 ⇄ ⚙️

Select a scenario or edit the JSON-RPC payload, then click "Dispatch MCP Request" to execute live.

16 Exposed MCP Tools

All available tools exposed via tools/list and callable via tools/call.

Tool Name Description Input Parameters
agent_message Sends a message to an internal sub-agent (sales, seo, booking, content, reporting) and returns its response. { toAgent, message, fromAgent? }
audit_website Fetches a live URL and returns real on-page SEO signals (title, meta, H1, word count, alt text, viewport). { url: string }
get_services Returns the business service catalog: packages, starting prices, and delivery timelines. {}
create_lead Saves a qualified sales lead and triggers webhooks, HubSpot sync, and Twilio SMS. { name, email?, phone?, budget?, notes? }
check_availability Fetches live upcoming discovery call booking slots from Calendly. { days_ahead?: number }
book_discovery_call Generates a single-use Calendly scheduling link for a client. { name, email }
create_calendar_event Creates a calendar event directly on Google Calendar. { summary, start_time, end_time, attendee_email? }
create_square_payment Generates a Square payment checkout link with line items and amount. { title, amount, currency? }
create_invoice Generates a QuickBooks customer invoice with line items and due date. { customerName, customerEmail, items }
create_task Creates a task in ClickUp with priority and description. { name, description, priority? }
save_content Saves drafted blog copy, headlines, or meta tags for team retrieval. { title, content_type, body }
get_recent_leads Reads captured pipeline leads for reporting and summaries. { limit?: number }
get_website_analytics Fetches traffic metrics (sessions, users, pageviews) from Google Analytics (GA4). { days?: number }
get_search_rankings Pulls top queries, impressions, and CTR from Google Search Console. { days?: number }
get_seo_domain_overview Semrush domain overview (organic keywords count, traffic estimates). { domain: string }
get_backlink_summary Ahrefs backlink profile overview (referring domains, live backlinks). { target: string }

Client Configuration & Integration

Plug this server directly into Claude Desktop, Cursor IDE, or custom code.


    

🌐 Universal MCP & Agent Discovery

Clients can query /.well-known/webmcp-manifest.json for the WebMCP manifest, or /.well-known/agent-card.json for the decentralized agent card.

WebMCP Manifest ↗ Agent Card ↗