"k6" matching MCP servers:

Local-first agent memory & knowledge service: hybrid retrieval (vector + BM25/RRF) + memory governance (decay, dedup, freshness), REST + MCP dual protocol, fully offline without LLM. Python.

MCP server that enables generating songs or instrumental music using Suno V6, V6_MINI, and V6_WILD models via the Kie.ai API, with support for checking task status and retrieving audio URLs.

Enables MCP clients to manage LLM inference memory through virtual paged-attention KV-cache block mapping, non-contiguous physical page allocation, zero-copy fragmentation tracking, radix-trie prefix caching, INT8 quantized compute, and speculative decoding verification. It also exposes prefill and decode latency telemetry so edge deployments can be benchmarked and tuned without external dependencies.

Enables edge AI agents to optimize inference memory and latency by mapping logical blocks to non-contiguous physical pages, tracking fragmentation, and exposing related MCP tools.

Sends ideas to a locally running LM Studio model for adversarial review, enabling critical feedback and devil's-advocate interrogation of proposals.

Enables to build and query a knowledge base with retrieval-augmented generation, supporting document ingestion, hybrid search, and live data integration from external APIs via MCP tools.

Official MCP Server for GPT-5.6 Sol Coding & Reasoning API. Connects extended reasoning capabilities directly into Claude Desktop, Cursor, and Windsurf.

A proxy that intercepts MCP responses to reduce token consumption by compressing them via a pipeline and exposing only two meta-tools to the LLM.

Enables MCP hosts like Cursor or Claude to query an enterprise knowledge base via stdio, providing RAG-based answers with cited sources.

A pure-AI cognitive extension that forces standard 3-Layer Validation to prevent hallucination in academic research.

Provides read-only access to local AI work center metadata including status, providers, usage, projects, and tasks via MCP.

Enables calculating PagedAttention KV cache allocation and GPU memory budgets for vLLM-style chunked prefill workloads via MCP.

Enables hot-swapping llama.cpp models in a running Claude Code session via launchctl or systemd, preserving conversation context across model changes.

A read-only MCP server that gives Codex and other clients a Kimi K3 second opinion through OpenRouter, enabling algorithm review, code review, and project planning.

MCP server for document intelligence via x402 micropayments. 6 tools: document analysis, invoice extraction, screenshot data, alt text, PII detection, sentiment analysis. Pay-per-use with USDC on Base — no API keys needed.

A local AI skill composer that reads, analyzes, compares, and combines SKILL.md skill files, enabling modular skill reuse like Lego blocks.

An MCP server that provides technical consultation, code review, and code explanation by integrating with OpenAI's Codex CLI, enabling AI-powered coding assistance in a sandboxed, read-only environment.

MCP server for the operbots panel, enabling AI assistants to manage tasks, Telegram bots, flow scenarios, conversations, knowledge base, and AI connections with the user's permissions.

Enables natural language image augmentation via the MCP protocol, using Albumentations to apply transforms like blur and rotation from plain English descriptions.

Local MCP server for urban analytics that wraps BlocksNet agent, enabling natural language urban questions to be answered with structured JSON results.