OOretz Factory routes every prompt through a multi-provider gateway. Hot-swap between Claude, GPT, Gemini, and Mistral per-message. Different models excel at different tasks — we route automatically, you can override.
Each request runs through a 3-layer routing decision before hitting a model:
Free → Haiku, Pro → Sonnet, Enterprise → Opus. Every user gets a sensible starting point.
TypeScript prefers Claude. Python prefers GPT. Java/Kotlin prefers Gemini. Rust/Go prefers Mistral. Caller can override.
When the manifest has 10+ entities or complex workflows, we bump to the next tier up. Pro user gets Opus for free on hard prompts.
Fast + cheap for simple TypeScript/React
Default model for the free tier. Excellent for simple iterations under 30 entities.
Best balance of quality + cost for complex apps
The workhorse. Most production OOretz Factory builds use this.
Top-end reasoning for complex architecture
When the prompt has nuanced business rules or multi-step workflows.
Cheap Python + data work
Cheapest model in the gateway. Great for simple Python or data work.
Strong Python + general-purpose code
Solid all-rounder. Often used as the second voice in ensemble voting.
Top-tier reasoning across all surfaces
Use when Opus struggles or when you need OpenAI-specific strengths.
Massive context, fast Java/Kotlin
Highest context-to-cost ratio. Good for large existing-codebase tasks.
2M context for huge multi-file work
2M context window unmatched. Use when 1M is not enough.
EU-hosted, strong Rust/Go
EU-hosted option for data residency. Strong on systems languages.
New models added quarterly. Existing apps re-route automatically when better/cheaper options appear in the same tier. Override per request via API or operator UI.
Start free →