⚠
Heads up: input & cache figures are best-effort while we're in beta. If they don't match your harness's actual usage, please report it to us on Discord.
Your lane
Base URL
Model
API key••••••••
Usage · your run
Requests
—
total
Input
—
Cached
—
Output
—
generated
Cache rate
—
of input warm
Value · DeepSeek V4.1 Flash pricing
Off-peak value
—
cache-inclusive · metered by LiteLLM
Peak value
—
at standard rates
Uncached · $0.15/M—
Cached · $0.003/M—
Output · $0.60/M—
Total (off-peak)—
Updates in real time as your lane processes requests · Singularity Reserved Inference
Admin · all users
—
Users
—
Approved
—
Waitlist
—
Total off-peak $
—
Keys in pool
—
Key requests
Key pool · paste LiteLLM keys, one per line
| User | Status | API key | Reqs | Input | Cached | Output | Off-peak | Peak | Actions |
|---|
Approve users, assign their API key, and trace spend · refreshes on load