Self-hosted general-purpose agent with offline-first reasoning, RAG, and MCP integration for enterprise workflows.

Feature-rich interface enabling image and video generation with flexible outputs, robust retries, and detailed progress tracking, tailored for WaveSpeed integrations.

Drop-in server delivering private, local SDM-based confidence estimates for LLM pipelines with adaptable training and tooling.

MCP Server integrating MCP Clients with Stability AI-powered image manipulation functionalities: generate, edit, upscale, and more.

AI-optimized author disambiguation using OpenAlex with robust affiliations and AI-friendly outputs.

Real-time vision-enabled integration delivering AI-driven screen analysis, streaming, and UI interaction across Windows, macOS, and Linux.

Local MCP server for semantically storing and retrieving text passages using Ollama embeddings and ChromaDB. It enables memorizing single or multiple texts, and contents from PDFs, with conversational chunking to create meaningful memory chunks. Retrieve results by meaning rather than keywords, supporting privacy-preserving, offline knowledge management.

Bridges MCP servers with OpenAI-style LLMs, translating tool specs into function calls and routing results bidirectionally across cloud and local endpoints.

Java SDK with Spring integration enabling standardized model tooling, transports, and async communication for seamless AI interactions.

Bridges Ollama with MCP to deliver local LLM control, OpenAI-compatible chat, and registry-based model management with privacy.

Gradio-based client demonstrating dual STDIO and SSE communication, with dynamic tool loading and debugging.

A type-safe, YAML-driven multi-agent framework with ACP and MCP support for scalable LLM-powered workflows.

Async MCP server enabling Claude Desktop and compatible clients to generate images, videos, and music using Fal.ai models via multiple transports.

AiCore now includes a comprehensive observability module that helps you track, analyze, and visualize your LLM operations. Now with MCP support!

Python MCP client enabling seamless LLM interactions via mcp.run with broad provider support and optional local/offline options.

ModelContextProtocol.jl

End-to-end automated ML workflow: data analysis, preprocessing, model selection, tuning, and visualization with Claude Desktop integration.

Unified evaluation and experimentation interface leveraging Patronus to run single/batch evaluations and data-driven experiments with remote evaluators.

A lightweight client enabling OpenAI-compatible access to local inference backends via configurable tools.

A lightweight bridge to access ClearML data—experiments, models, artifacts—with real-time metrics and cross-project search to enrich AI-assisted conversations.

Offline-first AI framework delivering Dockerized full-stack apps with declarative PKL and local LLMs.

Trusted runtime leveraging IPFS immutable storage and WASM sandbox for decentralized tool hosting and execution.

Modular autonomous intelligence with agentic networks, memory, multi-modal generation, and blockchain-enabled skillware.

Cross-platform LLM client enabling multiple providers, local inference, and modular tooling for flexible AI workflows.