Platform

Products

Private AI infrastructure — from compute to agents.

Target groups

Use Cases

For enterprise, SMBs, and individual developers.

Knowledge & Support

Resources

Everything you need to succeed with Mycelis.

Enterprise

AI at scale.
Costs under control.

AI costs can spiral quickly when every request hits your most powerful — and most expensive — model. Mycelis smart routing directs each request to the most cost-effective model that can handle it, without sacrificing response quality.

Right model, right price, every time

At scale, even small differences in per-token cost add up fast. Mycelis lets you configure routing rules that send simple classification or summarization tasks to fast, cheap models while reserving your high-capability models for tasks that actually need them. The result: the same output quality at a fraction of the cost.

What you get

Smart Model Routing

Define routing rules in your agent configuration. Route by complexity, task type, or keyword — automatically sending requests to the most cost-efficient model that fits.

Usage Analytics

See token consumption, cost breakdowns, and usage trends per workspace, agent, and model. Identify which agents or users are driving cost before it becomes a problem.

Mix Dedicated and Commercial

Combine self-hosted open-source models for high-volume tasks with commercial models for quality-critical requests — all through the same gateway endpoint.

Budget Controls

Set spending limits per workspace or user group. Requests above the threshold are automatically blocked or routed to cheaper fallback models.

Frequently Asked Questions

How does smart routing decide which model to use?

You configure routing rules in your agent settings. Rules can be based on prompt keywords, request metadata, or a round-robin / cost-optimized strategy. Mycelis evaluates each incoming request against your rules and forwards it to the appropriate model.

Can I see costs broken down by department or project?

Yes. Because each workspace has its own usage tracked separately, you get cost visibility at the workspace level. Use one workspace per department or project to get clean per-team cost reports.

What happens when a budget limit is hit?

You configure the behavior — requests can be rejected with an error, silently routed to a cheaper fallback model, or trigger an alert to the workspace admin. The default behavior is configurable per workspace.

Control your AI spend from day one.

Create a free account and set up smart routing in minutes.

Get Started Free