discode.ai

100+ AI models, one interface. ECO friendly.

Productivity

SaaS

Artificial Intelligence

Visit WebsiteSee on Product Hunt

@peterbuch it looks to me like this is a chat bot product. Is there a way to use discode via API (thinking about connecting Pi Coding Agent to it)?

Eco-friendly angle is genuinely underrated in the AI space - most people don't think about the carbon cost of hitting 5 different APIs vs routing through one. How does the model selection work under the hood? Does it pick the model for you based on the task, or is it always manual?

The Eco readout is the part I haven't seen anyone else actually ship, and it's overdue. What I keep wondering is where the CO2 and water numbers really come from. Providers don't publish per-request energy or which hardware served you, so is this basically tokens times a per-model coefficient, or something more grounded? Reason I ask: if the estimate is off by 2-3x, the eco-slider could quietly nudge people the wrong way. Either way, putting that bill in front of people is the right instinct.

I saw this kinda products but why discode unique?

Looks like a great one, How you are deciding the best model to be used for a prompt?

Such a great initiative. Definitely feeling conflicting trying to keep up with all the incredible AI progress while being mindful of the resource strain.

The model-choice reason is the part I’d make very visible.

As a builder, I’d love a tiny “why this route” card that turns each answer into a learning loop: task type detected, constraints it cared about (privacy / latency / quality / eco), and what signal would have pushed it to a stronger model.

That would help users trust the router without needing to understand 100+ model names, and it also gives you cleaner feedback when the route feels wrong.

the auto-routing is the real differentiator here — most multi-model platforms still make you pick manually, which defeats the purpose if you don't already know which model handles what best. curious how the routing logic works under the hood, especially for edge cases where two models are equally good but one costs 10x less. the CO2 tracking is a nice touch too, haven't seen that baked into a router before.

That's the trap we hit when we tried wiring re-asks back into routing. Raw re-ask rate was a noisy label, roughly half of ours were the user refining their own question, not the tier failing them. We had to gate on intent: only count a re-ask as an escalation signal when the follow-up keeps the same semantic intent as the original, otherwise a chatty user reads as a broken router. Worth deciding that filter before re-asks ever touch the tier boundaries.

Congrats on the launch! Which sources do you actually use to calculate CO2, water and energy?

Austria 👋 Building AI in adjacent space (DTC ad generation) and the multi-model routing point lands hard. Used to force everything through one pipeline early on, outputs were always mediocre at one step. Different tasks need different models.

The Eco angle is the part nobody else is showing. Most users don't know that a 1-line prompt to GPT-5 costs more than 50 to a smaller model. Curious, do users actually shift behavior when they see the readout, or is it more of a conscience check?

Really interesting approach to multi-model access. One thing I've noticed working across different AI models is how differently each one "knows" about specific brands or industries — the variance between ChatGPT vs Gemini vs Claude on the same query can be surprisingly large. Congrats on the launch, curious how you're handling response inconsistency across models!

Very curious how you determine the auto routing. Are you just using publicly available benchmark data? What is your criteria for which of the 100+ is selected? And how will you keep it up to date as new models are released?

Cool, it looks lime something I've done on Codex, to setup multi model for every sub agent. discode ai do this thing natively. I think it can save more token and get faster response with simple or complex chat. I'm cursious is discode has a main model to judge and distribute which model should do what?

100+ models behind one interface is a bold scope tbh, curious how you keep the UX from feeling overwhelming when there's that much choice. also the "eco friendly" angle is interesting, what's actually driving that — smarter routing, less wasted compute, or something else.

The CO2 footprint tracking per prompt is a genuinely fresh angle -- most AI tools pretend environmental cost does not exist, but making it visible nudges users to think twice before over-prompting, which is a quiet but powerful design choice.

The eco footprint per request is a great touch. What data source are you using for the CO2 estimates per model? Numbers vary a lot by datacenter location and grid mix, curious how granular you can get.

Congrats on the launch! 🚀

The eco angle is refreshing, but I also really like the on-device privacy filtering. Most AI routers focus only on cost and speed, while privacy and model choice are just as important.

Curious how transparent the routing explanation is for non-technical users can they easily understand why a specific model was selected?

Hi Moriz and team; congratulations to the launch. I really like the one-device privacy filtering feature and the overall ECO dimension (amid the hottest days ever in our region😓) Good luck and keep going!