Launch of discode.ai: An AI router that also considers the environment

discode.ai routes AI requests to the most efficient AI model

With some AI models, it feels like taking a van to the ice cream shop around the corner – and nobody tells you that it could just as well be done by bike or on foot.

The Multi-LLM router discode.ai changes this. The ESG Cockpit team is very pleased to support this innovative project and present it here. The platform is now live and helps users apply artificial intelligence in a responsible and efficient way.

Screenshot of the discode app with three sections: top left, a login screen with a disco ball and the text "You choose the rhythm, not the algorithm – Ein Chat. 100+ KI-Modelle. discode wählt das richtige." along with a "Sign in with Google" button. Bottom left, the "Turntable" section with three sliders for the criteria Smarter, Speed and Eco, which users adjust to weight the AI model selection. Right, the "discode Crew" section featuring four comic-style characters (Gloria Greenor, Sven Heimdall, Veritas Jones and DJ Smarty McFly), representing the values Eco, Private, True and Smart. — Smarter, faster, or more sustainable? With the discode.ai turntables, users weight the criteria for selecting the optimal AI for their request. The four heroes of the discode Crew guide them through the process. They represent the values Eco, Private, True, and Smart.

Sustainability as a core criterion

discode.ai routes user requests to wherever they can be processed most efficiently. At the same time, it does not patronize users. The routing is based on a user-defined weighting of criteria such as speed, quality, and sustainability.

Discover discode.ai

The Eco-Score: sustainability at a glance

For each response, the Eco-Score shows the estimated environmental cost of generating the answer:

CO₂ (in grams) – caused by greenhouse gas emissions from electricity generation, especially from fossil energy sources
Water (in milliliters) – required for cooling AI data centers, which generate heat through computational workloads
Energy (in watt-hours) – estimated based on model complexity as well as input and output token count

Screenshot of discode.ai on the topic of the Eco-Score. At the top, a banner in synthwave/cyberpunk aesthetics showing a person in a meditative, trance-like state, enveloped in lighting effects, standing against a backdrop of plants. The banner includes the question "Was kostet deine Frage den Planeten?" (What does your question cost the planet?). Bottom left, the note "Stell eine Frage … und sieh, was die Antwort kostet" ("Ask a question … and see what the answer costs"), explaining that the Eco-Score shows the CO2, water and energy consumption of each answer. Bottom right, an overview titled "Was wir pro Anfrage auswerten" ("What we evaluate per request") with the three metrics CO2, water and energy, along with brief explanations of how they are calculated. — *The Eco-Score of discode.ai shows how much CO₂, water, and energy is consumed in generating an AI response.*

Collaboration with the ESG Cockpit team for AI impact analyses

discode.ai is continuously working on improving data quality. This also includes collaboration with the ESG Cockpit team. In a joint working group on AI impact analyses, the partners are developing more robust approaches to assessing the environmental impact of AI applications.

discode.ai routes AI requests based on speed, quality and eco-efficiency. The ESG Cockpit team provides support in the area of AI impact analysis.

1 Product of the Day and #2 SaaS of the Week (Product Hunt)

A special highlight: the Vienna-based development was awarded #1 Product of the Day and #2 SaaS of the Week on Product Hunt. Voting for the tool is still open:

Vote for discode.ai

Product Hunt is a well-known community platform where mainly tech and software companies showcase their new products.