Where it runs, who can reach it

Security & working with us

Where the system runs, who can reach it, what is recorded, and how our access ends are settled in writing before anything goes live. This page also says what it is like to work with us.

FIG. 1 · Access planIllustrative
Runs on
Your servers in Türkiye, cut off from outside networks
eqlx access
Named accounts your team issues, time-limited, logged
Remote access
None standing; opened by your team on request
Outside model services
None
At handover
Accounts closed, keys rotated

Where it runs

It runs where your data already lives.

Regulated work runs on your own servers, cut off from outside networks where required. Other work runs in your cloud account or with a local provider you choose. eqlx does not host your system.

Outside model services are never assumed; one is used only with your explicit approval. We settle where it runs, and size the hardware, in the Decision Blueprint.

Your data never trains anyone else's system.

Your data, model weights, prompts, memory, and decision history stay in your environment. A model that serves several clients is trained separately for each one.

Banking data stays in Türkiye where the rules require it.

Decision Blueprint

What it connects to

Integration is the base under all three kinds.

Part of every engagement, not a fourth kind of AI: it is how each kind reaches the systems your team already works in.

  • Core banking

    The accounts, limits, and payments a decision reads.

  • ERP and MES

    Orders, work orders, and the production line's own records.

  • CRM and the contact center

    Customer history, calls, and complaints.

  • Your decision engine

    Your rules and thresholds stay in it; the system feeds it.

  • Documents and data

    Document stores, data warehouses, and reports.

  • Your servers or cloud

    Where it all runs, inside your environment.

We build and integrate

We build the model, small language model, or agent, and connect it to your systems.

We integrate what you run

We connect what you already run: an existing model, decision engine, or product.

We come into your environment, build beside your team, and hand the system over running.

Access and permissions

We reach only what the work needs, and only while it needs it.

We work through accounts your team issues, with the least access the work needs, and every access is logged. There is no standing remote access: your team opens it when the work needs it, and closes it after.

An agent's tools are read-only by default. An agent never grants itself a tool, permission, or approval.

At handover, our access ends.

Handover comes with documentation, runbooks (step-by-step operating guides), and training, so your team runs and changes the system without us.

Our access is then removed, unless continuing support is separately scoped and written.

Three starting points, each a separate decision that stays yours: the Decision Blueprint, a validation build, then production, integration, and handover.

How an engagement runs

Weights and IP

What we build for you is yours.

Customer data remains customer data. Customer-specific deliverables belong to the customer under the engagement terms. Pre-existing and reusable eqlx technology remains eqlx IP. Third-party and base-model licenses continue to apply.

The trained weights of your model are a customer-specific deliverable. A written list of what belongs to whom is agreed before the work starts.

Records and audit

Every step leaves a record your auditors can read.

Each model version, test result, and agent step is recorded in your systems, so your audit and validation teams can trace a decision back to its inputs.

We prepare evidence for your independent validation; we do not provide it.

Working with a senior engineer

A senior engineer leads, and the system never depends on one person.

A senior engineer leads each engagement. We take a limited number of engagements at a time. Everything is documented and handed to your team, so the system never depends on one person. We sign an NDA before any detail.

Reporting an issue

Found a security issue? Tell us.

If you believe you've found a security issue in code or configuration provided under an eqlx engagement, write to security@eqlx.ai. We assess each report against the scope of that engagement. Acknowledgment and fix timing depend on the finding and are not guaranteed by this page.

This website

The site is served over HTTPS at eqlx.ai. Accepted form submissions go through code that runs only on our server to a managed database in Frankfurt with restricted access. No advertising or cross-site trackers are set up; measurement is limited to cookie-free page views.

Equalize the gap.

Bring one decision. Leave knowing the first step.

Bring your security team's questions too: where it runs, who can reach it, what is recorded, and how our access ends.

https://eqlx.ai/security · Printed copy; the website is the current version.