Great Studios Inc.
GreatRouter is built by Great Studios Inc., a technology company focused on making AI infrastructure simple, transparent, and accessible.
Our Mission
We believe AI should be a utility — as easy to use as electricity or internet connectivity. GreatRouter removes the complexity of navigating dozens of model providers by providing a single, intelligent routing layer.
Why GreatRouter
The AI landscape is fragmented. Every provider has its own API, pricing model, and capabilities. GreatRouter unifies them behind a single API with:
- Automatic routing to the best model for your task
- Transparent pay-per-use pricing
- Built-in logging, tracing, and cost tracking
- Global low-latency edge network
Our Values
- Transparency: Every cost, every route, every decision is visible
- Simplicity: One API, endless possibilities
- Reliability: Global infrastructure with enterprise-grade uptime
- Fairness: Pay only for what you use — no minimums, no surprises
Open Source
GreatRouter is open-source under the Apache License 2.0. The full source code is available on GitHub.
The license allows you to use, modify, and distribute the software freely. The only requirement is that you preserve the copyright and attribution notices, and include the NOTICE file that identifies GreatRouter as the source.
Infrastructure providers
Inference requests are routed through these backend providers depending on the model and task:
| Provider | Purpose |
|---|---|
| Cloudflare Workers AI | Direct binding for ~73 models including Llama, Gemma, Flux, Whisper, BGE |
| Cloudflare AI Gateway | Universal /ai/run REST API — the primary path for unified billing across all providers |
| Fal AI | Fal-hosted models routed through Cloudflare AI Gateway |
| HuggingFace Inference Providers | Community and research models served via router.huggingface.co |
| Suno | Music generation via Suno API |
How routing works
-
Classification: The auto-router classifies your prompt into a task type (text, image, video, audio, music, code, translation) using keyword patterns and, for low-confidence cases, a lightweight LLM classifier.
-
Scoring: Every model capable of the classified task is scored by quality, price per unit, and optimization mode (speed, quality, or cost).
-
Selection: The highest-scoring model is selected. If inference fails, the router falls through the ranked candidate list until one succeeds.
-
Caching: Identical requests hit the semantic cache (Vectorize + KV) to skip inference entirely when the same prompt and model combination has been seen before.
Contact
Great Studios Inc., a Delaware C-Corporation
1209 N Orange St, Wilmington, Delaware 19801, USA
Email: contact@greatrouterai.com
Ecosystem
GreatRouter is part of the Greatapps family alongside GreatStudios and GreatChat.