Open source and public engineering projects that show HSP Hub work with MCP servers, AI tooling, infrastructure automation and data workflows.
MCP server connecting AI assistants to Bitrix24 CRM, tasks, chats, products, users and business processes through a safe API layer.
Turns Bitrix24 into an AI-accessible enterprise system with controlled tool calls, OAuth and typed API boundaries.
Dual-mode MCP server for NotebookLM automation: notebooks, sources, notes, research workflows, audio overview and Enterprise API scenarios.
Connects research workflows to AI agents with source management, notes and enterprise-ready API automation.
MCP server for managing Coolify infrastructure from an AI client: projects, services, deployments and operational actions through a typed tool interface.
Exposes deployment and infrastructure operations to AI assistants while keeping tool schemas, validation and safety gates explicit.
MCP integration for Satu.kz operations: products, catalog updates and marketplace data management scenarios through an AI assistant.
Combines marketplace API work, catalog validation and AI-assisted product data operations.
MCP server for controlled SSH operations: host management, command execution, interactive sessions, SFTP transfers, tunnels and reusable templates.
Turns remote server operations into typed AI-callable tools while keeping credentials, logging and transport boundaries explicit.
Local document-to-Markdown converter for PDF, DOCX, PPTX, XLSX, HTML and image-based inputs using Docling.
Creates a repeatable document ingestion step for knowledge bases, RAG pipelines and structured internal documentation workflows.
MCP server for connecting Gemini scenarios to agent workflows and extending AI tools through a dedicated integration layer.
Bridges Gemini CLI capabilities into agent workflows for coding, extraction, review and multimodal automation.
Small local RAG memory package for OpenClaw with a stdio MCP server, durable JSONL storage and an accompanying skill.
Gives an agent durable local memory without database or deployment overhead, with explicit tools for remember, search and forget.
File-based mode for long agent-assisted tasks: TODO, MEMORY, PROJECT-RESUME and STATUS-LOG preserve state between sessions.
Formalizes project memory and handoff for long-running AI-assisted work across sessions and agents.