MCP server development and integration

I build MCP (Model Context Protocol) servers and integrations that connect the data and systems of a company to AI assistants and agents, with permissions, caller identity and a record of every tool call. It is for teams that want Claude, ChatGPT or their own agents to work with real company data without handing the keys to a model.

An MCP server is a small program that offers functions to an AI model through a standard interface: read this record, list these invoices, create this task. MCP, the Model Context Protocol, is that interface. It is not the security. A company needs this work when an assistant has to answer from live data or act inside a system, and when the people responsible for that system must know which operations the model can perform, on whose behalf, and what it did.

What I build

  • A server for one system. A database, a CRM or an internal API, exposed as specific tools with defined inputs, returned fields, pagination and behavior on failure.
  • Remote gateways. A server on Cloud Run that cloud agents call over Streamable HTTP, with caller identity, secrets in Secret Manager and smoke tests. It can put many tools behind fewer, clearer entry points.
  • Public servers. LeanManufacturing.mx has a server with no key and no token: Claude, ChatGPT, Cursor or Codex can query its documentation and calculate OEE.
  • Open-source integrations. Airtable MCP lets an assistant such as Claude query an Airtable base in everyday language, installed through Smithery or the command line.

Questions to settle before a model touches your data

  • Which operations, exactly? A list of named tools, reads first. A general SQL tool needs far stronger isolation than an instruction that asks the model to behave.
  • Who is calling? The caller and the organization come from the authenticated context of the server, never from an identifier the model supplies.
  • What can the account of the server do? A database identity with narrow grants, and row-level policies where they apply. Parameterized queries stop SQL injection, but they do not decide whether this caller may read that invoice.
  • Where do the credentials live? In the secret store of the runtime, rotated and separate for each environment. Never in a prompt.
  • What happens on a write? Schema validation, business rules and authorization before execution, and for consequential operations an approval bound to the exact operation and its arguments.
  • What is recorded? For each call: a trace id, the tool and its version, the decision and the outcome.

How the work runs

  1. Inventory. I list the systems, the data in each one and the questions and actions the assistant should handle. A short list of tools comes out.
  2. Contracts for the read tools. Each tool gets a written contract: identity, authorization, result fields, limits on time, size and rate, and what the audit keeps.
  3. The refused cases. For each valid request I test its denied counterpart: another organization, an expired credential, a forbidden field, an oversized result, a database that is down. Retrieved text is treated as untrusted data.
  4. Writes, later and with approval. Changes arrive once the reads are trusted, with idempotency keys for retries and approvals that expire.
  5. Two transports. Local stdio for the machine of a developer, Streamable HTTP for the cloud. On Cloud Run the caller is identified by the platform and again by a bearer token in the application.
  6. Smoke tests and evidence. Smoke tests cover health, auth, initialize and the tool list, and the ledger of tool calls can be exported for review after a run.

What you receive

The server in your repository and your cloud project, with its container and pipeline files, the tool contracts, the permission model in writing, the denial tests, the smoke tests and the steps to register it with your assistants and agents. The credentials stay in your secret store.

What this does not include

An MCP server is not a security audit. I do not expose a general SQL tool on a production database because a prompt asks the model to be careful, and a second model call is not a permission check. If the database has no roles or policies that match your rules, that work comes first.

Where the evidence is

MetaMCP is my open-source gateway. Its case study covers the path to Cloud Run: both transports, the manifests for Agent Registry, a Google ADK sample, the smoke tests and the export of evidence. Airtable MCP is a public integration between a database and an assistant. LeanManufacturing.mx runs the public server described above. The cards on this page state where each one stops.

Work and evidence

MetaMCP as a Cloud Run MCP gateway (opens in a new tab)

My contribution: Built and documented the cloud-facing gateway path, registration artifacts, smoke checks and evidence export.

What this supports: Public implementation account covering transport, caller identity, secrets and tool-call evidence.

What it does not establish: The case study is not an independent security audit or a published concurrency benchmark.

Source reviewed:

Airtable MCP (opens in a new tab)

My contribution: Built an open-source MCP integration for querying and editing Airtable bases.

What this supports: A concrete database-to-agent integration with a linked public repository.

What it does not establish: An integration example, not evidence that arbitrary corporate databases or tenant policies are secured by default.

Source reviewed:

Common questions

What is an MCP server and does my company need one?

It is a small program that gives an AI assistant or agent a defined set of operations on a system: read a record, list items, make a change. You need one when the assistant has to answer from live company data or act in your systems, and you want that access limited, attributed and recorded.

Is it safe to connect our database to an AI assistant with MCP?

It can be, when the limits are enforced outside the model. The server uses a database account with narrow grants, takes the identity of the caller from authentication and keeps queries apart from changes. The protocol alone does not give that protection. The design around it does.

Can Claude or ChatGPT work with our internal systems through MCP?

Yes. Clients such as Claude, ChatGPT, Cursor and Codex can call an MCP server. For a server in the cloud the transport is Streamable HTTP, and for the machine of a developer it is stdio. I build both paths.

Do you build MCP servers for companies in Mexico and the United States?

Yes. I work remotely from Monterrey with teams in both countries, in English and in Spanish, on contract or in a senior role. The server is deployed in your cloud account, so access to your systems stays under your control.

Do we need an MCP consultant or an engineer?

An engineer, if the goal is a server in production. I am a Senior AI Engineer who takes contract work: I help decide which systems to expose and which to leave alone, and I write the server, its tests and its deployment.

Related questions

Open to senior AI engineering roles and contract engagements, remote in the US, Mexico and Latin America.