Nyquest routes every request to the right model, grounds it with memory and live web data, and compresses what goes over the wire. The platform runs at app.nyquest.ai; the compression engine, the Claude Code plugin and the SDKs live here.
Keeps large tool results out of Claude's context: parks them on disk, leaves an exact-line digest, and lets Claude recall precise slices on demand. Measured on a 13-turn session: −57% cost, −52% final context, identical output. Free; no account needed for local mode.
/plugin marketplace add Nyquest-ai/nyquest-claude-marketplace
/plugin install nyquest@nyquest
| Repository | What it is |
|---|---|
| nyquest-claude-mcp | Context Manager plugin for Claude Code (hook + MCP server) |
| nyquest-claude-marketplace | Plugin marketplace: /plugin marketplace add Nyquest-ai/nyquest-claude-marketplace |
| nyquest-rust-fullstack-3.2.0 | Semantic compression proxy for LLMs, full Rust stack (open source) |