(Modified:
2026-11-09)
— Written by
SimeonOnSecurity— 7 min read
A case study of self-hosted Hindsight, observed local Qwen routes, and request-scoped GPU work without a hosted model subscription on the inspected paths.
(Modified:
2026-11-08)
— Written by
SimeonOnSecurity— 11 min read
Wire Hermes Agent and other agents to models you control through one private endpoint, route each request to the right local or cloud model, and keep keys and budgets under your own policy.
(Modified:
2026-10-21)
— Written by
SimeonOnSecurity— 10 min read
Give each AI agent its own broker key with rate limits, token and cost budgets, and a fail-closed model allow-list, so one runaway caller stops at its own ceiling.