Documentation
Everything you need to integrate cachly into your stack. Pick a topic below or jump straight to the API Reference.
AI Memory — 3-Layer SystemNew
Give your AI persistent memory. learn_from_attempts + recall_best_solution + auto-pilot via copilot-instructions.md. Setup wizard with provider config.
Recall BenchmarkNew
Open methodology + results: +33% Precision@1 vs BM25, ~94% fewer context tokens — reproducible, defended by a CI gate.
API Reference
Interactive OpenAPI / Swagger UI for the cachly REST API.
Cache Engines
Dragonfly, Valkey, and Redis – choose the right engine for your workload.
Semantic Caching
Vector-based similarity search with multimodal support (text, images, audio).
Cluster Mode
Horizontal scaling with Valkey Cluster – automatic sharding and failover.
Agents SDK
Purpose-built caching layer for AI agents, tool calls, and chain-of-thought.
Terraform Provider
Infrastructure-as-Code for cachly – manage instances, keys, and tiers via HCL.
MCP Integration
Connect cachly to Claude, Cursor, Windsurf, Copilot and any MCP-compatible AI editor. Provision instances and cache LLM responses — in natural language.
IDE PluginsNew
VS Code & IntelliJ plugins — see what your AI learned, recall counts, and estimated token savings. One-click install.