Examples

Everything runnable lives in the open-source repo at github.com/Mindola-ai/mcp-server. It holds the public docs and example code for the REST API and the hosted MCP server.

Worked use cases

Ten end-to-end examples in the repo’s examples/use-cases/ folder. Each walks from problem to live lens: a prompt to paste, the conversation to expect, the tools involved, and the final result.

• A bike repair shop owner turns the shop’s pages into a website chatbot for price and opening-hours questions
• A long research report becomes a PDF chatbot clients can query directly, with citations
• A leadership coach builds an ask-me-anything page that answers from their method and captures leads
• A cooking channel creator answers viewer questions from their own recipes and episode notes
• A restaurant owner runs an always-on assistant for menu, allergen, and reservation questions
• A physiotherapist answers practice questions like insurance and referrals, with a firm boundary against personal medical advice
• A lawyer answers intake and process questions and qualifies leads, with a firm boundary against personal legal advice
• A real estate agent answers listing and bidding questions around the clock and catches the lead
• A startup founder gives investors, candidates, and users one AI page grounded in the deck and docs
• An agency owner runs an always-on pitch page and repeats the setup for every client

200 prompts

The repo holds two collections of 100 prompts each: docs/prompts.md has ready-to-use prompts organized by role and goal, and examples/100-prompts.md has one-sentence prompts grouped by industry. Three samples to paste into a connected assistant:

Create a Mindola lens for me. I'm a freelance UX designer with 10 years
of experience in fintech and healthcare. Interview me briefly to fill
the gaps, show me the profile, and only create it after I approve.

Turn my CV into a Mindola lens that answers recruiters' questions.
Here is my CV: [paste CV]. Keep the tone professional and add a
boundary that it never discusses salary expectations.

Create a Mindola lens in Dutch (locale nl) for my bakery. It should
answer questions about opening hours, allergens, and custom orders
from the text I provide.

Runnable scripts

The curl, Node, and Python scripts call the REST API, so they need an API key. Create one in the app (open Settings then API & keys), then export it:

export MINDOLA_API_KEY=mnd_your_api_key_here

• examples/curl/create-lens.sh creates a lens seeded with one text note and one link, and prints the JSON response
• examples/curl/add-knowledge.sh adds two more knowledge items to an existing lens; pass the lens id from the create response
• examples/node/create-lens.mjs creates a lens, polls every few seconds until all knowledge finishes ingesting, then prints the share URL; Node 18 or newer, no dependencies
• examples/python/create_lens.py does the same using only the Python 3 standard library

Client configs

The examples/clients/ folder holds ready-to-copy configs that connect an editor to the hosted MCP server: Cursor, VS Code, Codex CLI, Windsurf, and Gemini CLI. Setup steps for every client, including Claude and ChatGPT, are in the Connect overview.

The client configs and the MCP smoke test need nothing at all, because the hosted MCP server requires no authentication. To see a finished lens before you build your own, try the live demo at mindola.ai/me/charles-darwin.