4 tagged cost-reduction, measured the same way as everything else here.
MCP server Claude CodeCodexCursor +2
MCP server and proxy that compresses LLM prompts, tool output, and replies to cut token cost. Runs locally from the @llmtrim/cli npm package.
MCP server Claude CodeCodexCursor +2
Local-first vision-language pipeline inspired by VL-JEPA. Compress images, text, conversations, and RAG documents locally via Ollama before sending to any LLM API. Deterministic, fact-preserving token optimizer plus MCP server, FastAPI server and video processing. Runs locally from the latent-gate Python package.
MCP server Claude CodeCodexCursor +2
Reduce Claude Code token usage by 70-90% with cross-session memory and four-layer optimization: semantic codebase indexing, PostToolUse tool compression, turn summarization, and session consolidation. MCP server powered by Ollama — zero API calls, runs en. Runs locally from the cctx-optimizer npm package.
MCP server Claude CodeCodexCursor +2
MCP server "tokensaver-mcp" as configured in pozii/tokensaver. Runs locally from the tokensaver-mcp Python package.