ODIN — Sovereign AI Orchestration Layer
ODIN — Sovereign AI Orchestration LayerSnake River AI / Layer 5 of 6The router that decides which AI handles each query.What it does: ODIN sits above all your AI providers — Claude, OpenAI, Gemini, Groq, Together, HuggingFace, Ollama, LM Studio. It reads each prompt, classifies it, picks the right provider, forwards through the stack, and returns the response with full audit trail.Cost-aware: routes cheap queries to free local Ollama, expensive reasoning to paid providers. Cut your AI bill by routing intelligently instead of paying every provider for every query.Sovereignty-aware: prompts tagged sensitive stay local. Operator defines routing rules in plain YAML. No cloud dependency required.Composition: detects MIMIR (truth audit) and SAGA (continuity audit) on local ports. Operator opts in per rule via chain: directive. ODIN never silently rewrites your request path.What it does NOT do: No telemetry. No phone-home. No data harvest. - No vendor lock-in. Add or remove providers freely. - No subscription stack. Pay once per year. 8 providers in v0.1: Ollama (local, free), LM Studio (local, free), Claude API, OpenAI API, Gemini, Groq, Together, HuggingFace.Visual canvas (drag-drop routing UI, Business+ tier) ships in v0.1.5.Composes with: FREYA (Body), BRAGI (Voice), MIMIR (Mind), SAGA (Senses).Snake River AI — six layers that make any AI actually work.Install: Run ODIN-Setup-v0.1.0.exe. Settings UI at 127.0.0.1:7700/settings. No phone-home. Local-only.Requirements: Windows 10/11, Python 3.11 embedded runtime bundled in installer (~7MB exe). Optional: Ollama on port 11434, LM Studio on port 1234 (auto-detected).Sovereignty: All data stays on operator's machine. Audit ledger DPAPI encrypted at rest. License JWT validated locally.
Get it → clintwave84.gumroad.com