What is the routing matrix?
Cross-product of workloads and inference providers with recommended model tiers, policy notes, and o10 control-plane placement.
120 workload×provider combinations.
o10
Routing matrix by workload
Use case × provider pages show how o10 routes above each venue with unified evals, policy, and ledger.
Routing matrix pages show how o10 routes each enterprise workload above every provider. Eval-gated, with policy and ledger.
Routing120 guides
Support Assistant
RAG Summarization
Code Assistant
Batch Classification
Fraud Detection
Clinical Summarization
Knowledge Search
AI Agents
Real-Time Classification
Document Summarization
Translation
Data Extraction
Content Moderation
Recommendation Copy
Start hereQuick overview
How to use this index
Why separate from the routing hub?
The /routing hub covers concepts. Matrix pages are workload-specific playbooks per provider. Model tiers, policy notes, and rollout steps.
Each page links to integrations, savings, and model catalog.
How does o10 sit above providers?
Providers supply API access; o10 enforces eval-gated model selection, budget envelopes, and immutable ledger across all venues in one path.
Shadow mode before enforce on every rollout.
FAQFrequently asked questions
Common questions
How many routing matrix pages?
120 pages across 15 workloads and 8 provider venues.
Does this replace integrations pages?
No. Integrations explain stack placement. Routing matrix pages are workload-specific routing playbooks per provider.
How are recommended models chosen?
Candidate mini or open-weight tiers that typically clear evals at the use-case floor. Always verify in shadow mode on your traffic.
o10Set the envelope. o10 holds it.
See what you're overpaying.
Paste a week of traffic. Get the number that books the audit.
See what you're overpaying →