Judge visualization value and turn chart evidence into machine-learning actions.

An Agent-Computer Interface (ACI) for AI-driven machine learning.

AI/ML production readiness auditing. A remote MCP server that audits machine-learning projects for the issues that cause production incidents.

Enterprise-grade Machine Learning experimentation control plane for AI agents via Model Context Protocol (MCP). Train, validate, and diagnose 7 tabular architectures (LightGBM, XGBoost, CatBoost, RF, SVM, MLP) with zero arbitrary code execution.

McPortfolio: A Model Context Protocol server providing 9 specialized tools for LLM-driven portfolio optimization using natural language, covering mean-variance to machine learning approaches.

Machine-learning oral drug-likeness score for small molecules (SMILES in, 0–1 out), with an MCP server for AI assistants (Claude, Cursor, VS Code and other MCP clients). J. Chem. Inf. Model. 2026.

ML Referee evaluates whether candidate features add predictive value beyond a baseline for machine learning and forecasting. AI agents can preflight data, request quotes, and run authorized evaluations through MCP. Invite-only beta. Website: MCP endpoint: https://api.mlreferee.com/mcp

Language-learning coach MCP server for Claude Code: Claude writes the lessons, LangCoach keeps your weekly plan, streak, XP and spaced-repetition (FSRS) reviews on your machine.

Continual learning MCP server for AI agents — topology-aware knowledge graph with kappa-routing, exposed as 5 loop-phase machines (retrieve/route/act/learn/consolidate), causal metadata, and neural embeddings. One of four MCP servers in the [&] three-prot

This project demonstrates how to build an MCP server that bridges AI assistants with data analysis and machine learning capabilities.

A multi-layered cybersecurity platform combining machine learning, real-time Gmail analysis, and community intelligence to detect phishing across URLs, emails, documents, and messaging platforms.

A local Model Context Protocol server for training small text classifiers with Extreme Learning Machines (ELMs). It supports in-memory inference and optional SurrealDB persistence, prediction monitoring, and vector storage.

Media discovery platform built by harvesting open datasets, applying machine learning text classification, and implementing Model Context Protocol (MCP) agentic RAG guardrails.

Local-first AI planner that learns your preferences and turns events into memories. Your data stays on your machine.

A dedicated MCP (Model Context Protocol) tool for github, The MCP (Model Context Protocol) Tool for GitHub is a Python-based solution designed to help manage and track machine learning model context directly within GitHub repositories.

Static security scanner for machine-learning model files. Checks PyTorch, pickle, joblib, GGUF, SafeTensors, ONNX and Keras checkpoints for code execution before you load them. CLI, GitHub Action, load guard, MCP server.

A Model Context Protocol server with machine learning engineering tools: BPE token counting and cost estimates, BM25 search, paired bootstrap and McNemar evaluation comparison, JSON Schema validation, and prompt-injection and PII scanning.

Manage the AI skills you already installed for Claude Code, Codex and Antigravity: route each request to the right skill, keep your corrections as learnings, and benchmark skills on your local machine.

A simple yet powerful local AI Agent assistant that runs entirely on your machine. Option for remote LLM. Built for learning and deploy, Little MCP combines the power of open-source LLMs with advanced RAG that work with your personal documents. Included tools: real time weather, local documents RAG, local SQL database. Easy deploy your own agent.

## Contextual — Persistent Memory for AI Coding Agents AI coding agents forget everything the moment your session ends. **Contextual** fixes that by giving any MCP-capable AI client — for developers who use multiple AI coding tools at once — a real, persistent memory of your codebase: not just what the code says, but what changed and why. It's a **local-first MCP server**: your code, your index, and every query stay on your machine. The only network call is a lightweight, infrequent license check — no code, ever, leaves your device. Under the hood, Contextual builds a real understanding of your repository, not a text-similarity guess: - **Structural search** — tree-sitter AST parsing + call-graph resolution, so agents reason about how your code actually connects. - **Bitemporal history** — tracks not just when your code changed, but when the AI learned about it. Ask "did this bug fix happen before or after the agent's understanding of this function?" and get a real answer. - **Git-aware retrieval** — temporal, history-grounded context instead of a stale snapshot. - **Fast, self-contained** — index a 50k-line repository in about 5 minutes, then query it in ~94ms median, from a single ~0.8GB install. This extension connects Claude Desktop (or any MCP-capable client) to a Contextual installation already set up and activated on your machine — it does not provision Contextual from scratch.

**Persistent memory for AI agents that you can actually read.** Kybase stores what your agent learns as plain Markdown notes with `[[wikilinks]]`, searched by keyword and by meaning. It runs locally: notes live in an embedded Postgres on your machine (`~/.kybase`), with no Docker, no server and no account. **What your agent gets (18 tools)** - **Search** — `search_notes`: hybrid keyword + semantic search that returns the section each excerpt came from - **Read** — `get_note` (whole, windowed or one section), `list_notes`, `list_tags`, `list_folders`, `indexing_status` - **Write safely** — `create_note`, `append_to_note`, `replace_in_note`, `update_note` with conflict protection, `delete_note` / `restore_note` (trash) - **Link** — `get_backlinks`, `get_neighbors`, `get_graph` over the wikilink graph - **Organize** — `create_folder`, `update_folder`, `delete_folder` **Setup** Keyword search works right away. For search by meaning, pick an embedding provider in the settings: local **Ollama** (`ollama pull embeddinggemma`) keeps everything on your machine; **Google** or **OpenAI** need an API key and send note text to that provider. **Your data stays yours** `kybase-mcp export vault.zip` writes every note out as plain Markdown, folders as directories, readable in any editor, Obsidian included. **Want a UI?** The full Kybase app adds a web editor, a knowledge-graph view, share links, and one memory shared by Claude Code, claude.ai and Cursor: [github.com/Kyrzin/kybase](https://github.com/Kyrzin/kybase) Open source, AGPL-3.0.

**Run GPU workloads, deploy AI models, access inference APIs, and execute AI agents using Badgr's MCP server.** Badgr is an AI compute and workload execution platform for developers, coding agents, and AI assistants. Run workloads on cloud GPUs or compatible customer-owned GPUs (BYO GPU), with cost controls, execution logs, and verified results. **What can Badgr do?** - **GPU Compute / AI Workloads:** Run Python scripts, machine learning jobs, batch inference, and supported ComfyUI workflows. - **AI Model Deployment:** Deploy and serve LLMs using compatible runtimes such as vLLM and SGLang. Get a verified, OpenAI-compatible inference endpoint. - **AI API / LLM Inference:** Make AI API requests with response validation, supported fallback routing, and bounded execution. - **AI Agent Execution:** Launch coding and testing agents, run tasks, check acceptance criteria, and retrieve execution receipts. **Key features:** Cloud GPU execution, BYO GPU support, model serving, AI agent jobs, workload verification, execution logs, cost limits, and deployment management. Connect Badgr to your MCP-compatible AI assistant to execute real workloads instead of just generating instructions. Get started: https://aibadgr.com **Requires a Badgr API key. GPU workloads and other paid execution may incur usage charges.**

Parametric price benchmarking for engineering and procurement trade-offs. Agentic first, runs on your data, stays on your machine.

Modern experiment tracking that follows you across servers, messengers, and LLMs.

Schematic capture for neural network architectures. Draw the model, get the numbers, generate the PyTorch.

Measurement-based EV battery SOH estimation, validated on unseen batteries, exposed as MCP tools to an LLM agent.

Persistent absorption engine for Claude. Feed it code, docs, and failures — it builds a local skill graph any Claude session can query. Exposed as an MCP server.

Linter for LLM fine-tuning datasets: catches broken chats and tool calls, context overflow, leaked keys, PII and eval contamination before you train. CLI, Python API, MCP server, GitHub Action.

Solvro internal MCP server

Enterprise-grade dynamic prompt management system with runtime prompt application, enhanced MCP server, constitutional AI, workflow orchestration, session management, metadata handling, and comprehensive testing framework

MCP (Model Context Protocol) layer for sktime - Registry-Driven for LLMs

Neo MCP server — bring Neo, the autonomous AI engineer for ML, LLM, GenAI & data workflows, into Claude Code, Cursor, Codex & other editors. Describe a task; Neo builds it and writes results to your repo.

A comprehensive Python library for scientific computing and data analysis

A Model Context Protocol (MCP) server for agentic retrieval of financial market data

Periodic user-access-review (UAR) campaign runner

MCP server for intelligent news article categorization using embeddings and clustering

Lightweight Model Context Protocol servers for real-world actions: File Manager, Python Runner, URL Scraper, and Shell Exec.

Auto Dataset Cards / datasheets with Croissant + provenance

Batteries-included local RAG pipeline — ingest, index, serve

A beginner-friendly, production-aware ML workflow layer for students, analysts, and AI agents. Train, compare, explain, export, and deploy scikit-learn models without the boilerplate.

LLM hallucination & grounding auditor for RAG systems

Runtime allowlist and policy for agent tool-calls

Versioned prompt / template registry with A/B and rollbacks

A powerful MCP server for AI-driven Jupyter Notebook management and execution

A Python toolkit — AI for Science

Cross-browser tab grouping extension with secure MCP server and optional ML microservice.

A Proof of Concept MCP server for macOS, enabling AI agents to interact with macOS apps

RAG corpus poisoning detector — embedding anomalies, backdoor triggers

Chainable encode/decode/transform pipeline (base64/hex/rot/xor/url/gzip)

Squawk for backtests — static leakage/lookahead-bias detector for trading/ML strategy code, exposed over MCP.

Local model router / proxy across Ollama, vLLM, and cloud with fallback

Lint, version, and test prompts as code with a CI gate

MCP-native modular ML workflow engine — YAML-config-driven, DAG-orchestrated data analysis & model training, exposed as MCP tools

A full-featured MCP server wrapping the Kaggle API — 52 tools across competitions, datasets, kernels, models, benchmarks, and discussions

Classify findings against Microsoft's AI-agent threat taxonomy with mitigations

MCP server for spatial transcriptomics analysis through natural language interfaces.

Discover and map agent-to-agent / MCP communications and flag shadow AI

Durable, model-agnostic long-term memory for AI agents — stdlib, SQLite, MCP-native

Lint, curate, and prepare computer-vision datasets from your AI assistant MCP server.

MCP server for Google Vertex AI — 181 tools covering all v1 REST API endpoints