Product
Control agents and inference
in one place
Inferix sits in front of owned models and providers. See every call, route by policy, catch quality drift, and retrain. A closed loop, not a second stack of charts.
Why Inferix
Observe → Trace → Route → Drift → Retrain. Five products. One spine.
- 01
Observe: See every call
Latency, cost, tokens, and errors for agents and models. One dashboard, not five tools.
LensAI - 02
Trace: Follow the whole agent
Follow tools and model steps in one graph. Debug a bad run without guessing which hop failed.
TraceForge - 03
Route: Send each call by policy
Cheap owned SLM for simple asks. Strong path or provider when you need it. Policy decides, not a spreadsheet.
RouteIQ - 04
Drift: Catch quality before tickets
Alert when answers go bad versus a teacher or golden set. Know before users file tickets.
DriftWatch - 05
Retrain: Ship the next model safely
Feed FineForge. Ship a better model. Promote or roll back without guessing.
FineForge
One loop. Five jobs.
Day one you see traffic. Then you route, catch drift, and ship the next model.
LensAI metrics on every call.
TraceForge spans across the agent run.
RouteIQ policies across owned models and providers.
DriftWatch alerts when quality slips.
FineForge turns that signal into the next job.
Product suite
What each product does
LensAI
See every call
Latency, cost, tokens, and errors for agents and models, in one place.
TraceForge
Trace the whole agent
Tools and model steps in one graph, so you can debug a single run end to end.
RouteIQ
Route by policy
Send easy work to a cheap owned SLM, hard work to a strong path or provider.
DriftWatch
Catch quality drift
Alert when answers go bad against teacher models or golden tests, before tickets pile up.
FineForge
Improve and retrain
Turn drift signal into the next fine-tune. Promote or roll back model versions with confidence.
Who it's for
Teams that run models and agent fleets
If you already proxy LLMs or collect traces, Inferix ties routing, quality, and retrain together. Owned models and agents share the same telemetry and policy.
Try it locally
Walk the sales console, or follow quick start for a local Docker stack.