Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Ask GorillaDesk (gorilladesk-mcp) Who owes me money?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Read-only MCP server so Claude, ChatGPT, or Cursor can answer pest-business questions from your GorillaDesk account in plain English.
It looks things up. It does not create customers, change jobs, or take payments.
How many customers do I have?
Who owes me money?
What’s on the schedule this week?
Who are my technicians?
How’s money this month?
Give me a Monday morning briefing
Related MCP server: OfficeRnD MCP Server
GorillaDesk uses a company API key (Bearer token).
Open Addons → API
Generate / copy the key
Put it in .env as GORILLADESK_API_KEY (local) or paste it at /connect (hosted)
Docs: https://api.gorilladesk.com/docs/
cd gorilladesk-mcp
python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"
cp .env.example .env
# edit .env and paste GORILLADESK_API_KEYWire Cursor / Claude Desktop to scripts/run_mcp.sh (see mcp.json.example).
export GORILLADESK_MCP_HOSTED=1
export GORILLADESK_VAULT_SECRET='a-long-random-string'
./scripts/run_mcp_http.shPath | Role |
| Landing |
| Product page |
| Paste API key → mint bearer token |
| Streamable HTTP MCP |
| Health |
Client allows GET only
Tools annotated readOnlyHint=True
Owner tools return an answer string agents should read aloud
Product page lives under Relay at /gorilladesk. Marketing host: the relay-mcp Vercel project (fieldwork-mcp landing/).
pytest
ruff check src testsAnswer: "Do I have a customer named ___?", "Find customer Smith"
Search customers by name, company, email, or account number.
Inputs: query (required), limit. Returns: result
Fetch one customer by id. Prefer find_customer first.
Inputs: customer_id (required), include. Returns: result
Fetch one invoice by id.
Inputs: invoice_id (required), include. Returns: result
Fetch one job by id.
Inputs: job_id (required), include. Returns: result
Verify GorillaDesk credentials. Run this first if other tools fail.
Calls GET /v1/company. On success returns a small company summary.
Inputs: no named parameters. Returns: result
Answer: "How's business?", "Give me a snapshot"
Jobs + invoiced/collected for a date window (default 30 days).
Inputs: days, end_date, start_date. Returns: result
Answer: "How many customers do I have?", "What's my customer count?"
Counts active GorillaDesk customers and breaks down by status.
Inputs: no named parameters. Returns: result
Answer: "How much money came in?", "What did we invoice this month?"
Invoiced + collected totals for a date window (default 30 days).
Inputs: days, end_date, start_date. Returns: result
List/filter customers. Prefer find_customer / how_many_customers for owner questions.
Inputs: limit, offset, status, account_number. Returns: result
List/filter estimates.
Inputs: limit, offset, status, customer_id. Returns: result
List/filter invoices. Prefer who_owes_me_money / how_much_money for owners.
Inputs: limit, offset, status, end_date, +2 more. Returns: result
List/filter jobs. Prefer whats_on_the_schedule for owner questions.
Inputs: limit, offset, keyword, end_date, +2 more. Returns: result
List/filter payments.
Inputs: limit, offset, end_date, start_date, +1 more. Returns: result
List services. Prefer what_services_do_we_offer for owner questions.
Inputs: limit, offset. Returns: result
List users. Prefer who_are_my_technicians for owner questions.
Inputs: limit, offset. Returns: result
Answer: "What services do we offer?", "What's in the catalog?"
Lists GorillaDesk services.
Inputs: no named parameters. Returns: result
Answer: "What's on the schedule?", "How many jobs this week?"
Jobs overview for a date range (default 7 days).
Inputs: days, end_date, start_date. Returns: result
Answer: "Who are my technicians?", "Who's on my team?"
Lists GorillaDesk users with roles.
Inputs: no named parameters. Returns: result
Answer: "Who owes me money?", "What's my AR?", "Who has open invoices?"
Ranks open GorillaDesk invoices by amount due.
Inputs: top_n. Returns: result
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 19 tools
There is overlap between natural-language query tools (e.g., how_much_money, who_owes_me_money) and generic search tools (e.g., search_invoices, search_customers), though descriptions mitigate confusion with 'Prefer' hints. A few pairs like search_customers/find_customer and search_jobs/whats_on_the_schedule are nearly interchangeable and could lead to misselection.
The set mixes natural-language question phrases (how_much_money, who_are_my_technicians) with standard verb_noun patterns (search_customers, get_job). Each style is internally consistent, but the lack of a unified naming convention makes it harder to predict tool names across the whole server.
19 tools is slightly over the ideal 3-15 range but appropriate for the breadth of the GorillaDesk domain (customers, jobs, invoices, payments, estimates, services, users). The two-layer design (answer-oriented vs. data-access) adds functionality but also inflates the count.
The toolkit covers the core entities needed for a read-only business assistant. Minor gaps exist, such as missing get_* tools for payments, estimates, services, and users, but the search tools likely return sufficient detail for most 'ask' scenarios.
Ask Claude or any MCP client about your Larvabot leads and buyer-intent conversations, read-only.
Read-only MCP server for ClassQuill, a tutoring-business-management platform.
Google Search Console, GA4 and PageSpeed MCP server for Claude, ChatGPT and Cursor. Free, hosted.
Hosted CallRail MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
