Automate the repetitive work behind your online business.
I build custom data collection, monitoring, AI, and workflow automation systems.
Featured case study
TCG Market Intelligence & Automation Platform
A private system built to run an actual trading-card e-commerce operation: collecting marketplace data, normalizing products, tracking prices over time, and surfacing buying and selling opportunities.
System flow
- 1
Marketplaces
Data pulled from multiple marketplaces and APIs
- 2
Collection
Scheduled scraping and API ingestion
- 3
Normalization
Canonical product identification and deduplication
- 4
PostgreSQL
Historical pricing and market statistics
- 5
Analysis
Price trends, arbitrage and opportunity detection
- 6
Decisions
Buying and selling decisions, listings, automation
What it demonstrates
- Multi-marketplace data collection on a schedule
- Canonical product matching across inconsistent listings
- Historical price data stored in PostgreSQL for analysis
- Opportunity detection on top of normalized market data
- Built end-to-end as a solo developer
Built for my own e-commerce operation, not client work. It is not a public product. Sensitive credentials and third-party data are not shown.
Services
What I can take off your plate
Sold as outcomes, not technologies. Each system is built around a specific piece of repetitive work.
Data Collection
Stop copying data by hand. I build systems that gather it on a schedule, reliably and repeatedly.
- Website data extraction
- Public web data collection
- API-based data ingestion
- Browser automation
- Product & catalog collection
- Competitor monitoring
- Marketplace data collection
- Scheduled data gathering
Workflow Automation
Turn repetitive digital work into workflows that run themselves and notify you only when it matters.
- Repetitive browser workflows
- Spreadsheet & database automation
- Business process automation
- Scheduled workflows
- Notifications & alerts
- API-to-API integrations
- CRM & data synchronization
- Internal business tools
AI-Powered Workflows
Use LLMs where judgment is fuzzy and deterministic code everywhere else, so the system stays reliable.
- Classification
- Information extraction
- Summarization
- Semantic matching
- Research
- Categorization
- Natural-language processing
- Document processing
Data Systems
Give collected data a home so it can answer questions, not just pile up in a folder somewhere.
- PostgreSQL databases
- Historical data
- Data normalization
- Deduplication
- Change detection
- Data pipelines
- Dashboards
- APIs
Demonstrations
Automation you can actually look at
Demonstration projects and sample output. These show how the systems work, not client results.
Competitor Price Monitoring
Website monitoring- Sources monitored
- 8
- Products tracked
- 4,821
- New products
- 37
- Price changes
- 84
- Out-of-stock changes
- 16
Sample output shown for demonstration. Figures are illustrative, not client results.
Process
How a project runs
Fixed scope where it makes sense, so you know what you are getting before it starts.
- 01
Describe the problem
You explain the repetitive workflow and what it costs you in time or missed information.
- 02
Scope a concrete fix
I turn that into a fixed-scope proposal: what it collects, what it produces, and what it costs.
- 03
Build and verify
I build the system, you see it running against real data, and we fix whatever the data reveals.
- 04
Run and maintain
It runs on a schedule. I keep it working as sources change, with monitoring and support if you want it.
About
A developer who runs an online business, not a generic agency.
I started building automation because I was running my own trading-card e-commerce operation and kept hitting the same problem: too much of it was repetitive digital work that software should handle.
So I built it. Data collection, marketplace APIs, normalized product catalogs, historical pricing, analysis, and parts of the selling workflow. That system is still in active development and it is the honest reason I can build these systems for other people.
TungByte is how I extend that same capability to other online businesses. Small and technical by design: a direct line to the developer, with no middlemen whatsoever.
- •Deterministic code by default, LLMs only where judgment is genuinely useful
- •Real systems against real data, not slideware
- •No unnecessary infrastructure before it earns its place
- •You work directly with the person building the thing
Contact
Tell me what you are doing by hand
Describe the repetitive workflow and what it costs you. I will reply with whether it is worth automating and roughly what it would take.