Don't forget to be funny
We replaced our linear memory with three kinds of memory — semantic recall, permanent facts, and transient notes — and learned a lot about how to be funny.
architecturememorypgvectorgeminisystems-engineering
From the Team
The RidgeText Blog
Deep dives on how we build — AI architecture, mapping, and outdoor technology.
How we reduced LLM cognitive load with the Sub-Agent-as-a-Tool pattern
Piling 15+ complex tools into a single LLM was causing context confusion. Here's why we bypassed top-level routing, embraced the Sub-Agent-as-a-Tool pattern, and built a generic multi-agent backend that scales.
architectureagentic-designgeminisystems-engineering
How we cut our cloud bill 85% by splitting one service across two machines
A single always-on Cloud Run instance was costing us $160/month for compute headroom we only needed during LLM tool calls. Here's how we reshaped the system around Pub/Sub and a small dedicated worker — and what the latency numbers actually looked like.
architecturecloud-runpubsubcost-optimization
How One User Can Ruin It for Everyone Else - The Image Loophole
One User. Seven Days. 1,074 AI Images.
platformtransparencycredits
Mapping with In-Memory Layers to Reduce LLM Overload
How we built a Mapbox-compatible map compositor where the LLM orchestrates layers without ever touching GeoJSON — keeping context windows small and rendering deterministic.
mapboxllmarchitecturemapping
Ready to explore?
Start with a free plan and text RidgeText from wherever your adventure takes you.