Repository navigation
SlickToken is a platform for monitoring and planning LLM inference infrastructure. It answers capacity planning questions before you spend money on hardware: which GPU fleet handles your traffic, when you'll need to scale, what happens if you switch engines, and what the real cost of a multi-step AI workflow is.
Live data telemetry is optional and not required. Free until July 2027.
No data leaves your network.
Newest additions:
Updated Cost Calculation model fitted using real throughput measurements of 4 different GDDR cards
Basic portfolio management for compute markets (physical exposure with forward hedging computation)
Trend Incident storage
Load workload information from triggered incidents into the workload simulations
Cold starts accounting for autoscaling scenarios
Please reach out to sethk@slicktoken.ai for details.