Products
Web applications with accounts, permissions, and a workflow someone uses every day.
Custom software and AI
Custom software and AI development.
Web applications, APIs, distributed systems, data pipelines and cloud. LLM applications, agents, retrieval, automations, MCP servers and forecasting.
Software, close up
Workspaces, hardware and a sample screen. These are not client projects.
Sample interface. Not a client product.
A screen someone can use
Custom software is an interface and the system behind it. The panel is a sample, built for this page, so you can see the kind of screen. It is not a product we shipped for a client.
The work
Two lines of work, quoted per project. Scope and price are agreed in writing, in US dollars, before anything starts.
Web apps, APIs, distributed systems, data, cloud.
02AI developmentLLM apps, agents, RAG, automations, MCP, forecasting.
What gets built
Named plainly. Each line opens the page that describes it.
Software
- Web applicationsProduct interfaces and the application behind them.
- APIsContracts other software can call without guessing.
- Backend & distributed systemsServices, workers, and the coordination between them.
- Data engineering & pipelinesFrom an event to a number someone can trust.
- CloudAWS, GCP and Azure.
AI
- LLM applicationsA model doing a defined job inside a product.
- AI agentsA sequence of steps, with tools and limits set up front.
- RAG & knowledge assistantsAnswers from material you actually have.
- Automations & integrationsA repeated task, connected, and still stoppable.
- MCP serversOne system, exposed to AI clients on purpose.
- ML forecastingA model of a series you already measure.
Process
How an engagement goes. This is the working method, not a wall of reviews. The longer version is on the approach page.
- 01
You email us
Tell us what you need. We confirm scope and price in writing before any work starts.
- 02
You share what the work needs
Code, data, or the least access required. We don’t ask for your passwords.
- 03
We build, you review
Where AI tools are used, a person checks the work before you are asked to approve it.
- 04
You approve
Nothing is published or changed in your accounts without your approval.
Where this kind of system fits
Kinds of work, not a client list. No logos, and no claim that a particular company has hired us.
Operations
The software behind a repeated job: queues, workers, and a record of what already happened.
Knowledge
Assistants that answer from material you already have, with the source still visible.
How a case study would be written
The format is here. The clients are not. A real example is added only with written permission.
Tools used in the work
A working list, written as words. No logo wall.
Languages
- Python
- Node/TypeScript
- Go
- PHP
Data
- Kafka
- Snowflake
Cloud
- AWS
- GCP
- Azure
AI
- LangChain
- LlamaIndex
- MCP
Names, not endorsements. These are tools used in the work. They are not clients, sponsors, or partners.
What we promise, and what we don’t
Rules for the custom software and AI work on this site.
We will
- Agree prices in US dollars, in writing, before work starts.
- Keep your business facts as you gave them to us.
- Label AI-altered output where it matters.
- Stop emailing you the moment you ask.
We won’t
- Invent reviews, testimonials, features or results.
- Publish your work on this site without written permission.
- Promise rankings, sales or uptime we have not written into the scope.
- Use material you don’t have rights to.
- Add tracking pixels or cookies to our own emails or this site.
Working rule
Nothing about your business gets invented.
Questions first? That’s fine.
Email us with what you need. We reply during US business hours. hello@plusonets.com