Customised, Integrated Trading Operations
Choose Robokami to build and integrate trading operation systems tailored to your organisation's needs in short-term electricity markets.
How we workGet in touch
Tell us briefly about your portfolio and how your desk works; we will get back to you.
Your organisation, your system.
Robokami is set up around each client's own system and needs. It does not rip out the data sources, spreadsheets and routines you already have; it builds on them.
Data first: EPİAŞ Transparency, market feeds and your own generation and consumption records go into one order. Forecasting, positioning and order logic are then built on top, for day-ahead, intraday and balancing.
The algorithm does not replace the trader. The desk sets the rules, the system executes, and the reasoning behind every decision is logged.
Ready-made products.
Some parts no desk should have to write from scratch. We ship those as products: they plug into your own system, or run on their own.
- GIP2LoggerConnects to EPİAŞ GİP, processes the order book and trades in real time; a thin API layer for algorithmic trading.
- Forecast BoostCombines probabilistic generation and consumption forecasts with price forecasts; produces position and trade suggestions.
- Forecast HubCollects, versions and compares forecasts in one place. Which model or forecaster is how accurate becomes visible.
Open to agents.
Our products speak MCP as well as REST. Your own assistant or agent can query market data, log a forecast or pull a report, whichever model you use.
eptr2's MCP server and agent skills are open source. We help set up the same approach on your desk: tool schemas, permissions, audit trails and a human in the decision loop.
Working with agents.
There are three parties at the desk: the trader, Robokami and the agents. All of them look at the same data, the same tools and the same log. The agent gathers data, does the arithmetic and drafts; the trader decides; the system executes and stores the decision with its reasoning.
Your own agents can join this arrangement. Because our tools are exposed over MCP, your desk's assistant can ask Robokami questions and Robokami's agent can read from your system. Permissions are stated explicitly and every call is logged.
This is how we work ourselves: the data archive, daily checks, forecast monitoring and reports run with agents, and changes go through human approval. The setup we build for a client is the one we use.
For machines: llms.txt · agents.txt · content/
Resources
Open-source library, reports and newsletter.
eptr2
Open-source Python client for the EPİAŞ Transparency Platform: 200+ endpoints, a CLI and an MCP server. Apache 2.0.
Market reports
Intraday price and volume, imbalance cost, volatility: periodic reports and reporting for organisations.
Newsletter
Substack newsletter on the Turkish short-term electricity market (in Turkish).
# pip install eptr2
from eptr2 import EPTR2
eptr = EPTR2()
# day-ahead market clearing price (PTF)
df = eptr.call("mcp", start_date="2026-10-01", end_date="2026-10-07")eptr2 mcp-config --client claude-code # or vscode, cursor, claude-desktop
eptr2-mcp-server