Hexis Cloud

Product Introduction

1. Definition: Hexis Cloud is a hosted, enterprise-grade platform that functions as a centralized knowledge and tooling hub for human teams and their AI agents. Technically, it is a Model Context Protocol (MCP) server and a collaborative knowledge base that integrates with version control systems like Git.
2. Core Value Proposition: Hexis Cloud exists to solve the problem of fragmented and ungoverned AI agent deployment within organizations. It provides a single source of truth for validated company context, reusable AI skills, and secure tool access, enabling teams to scale their use of AI agents like Claude and ChatGPT with governance, security, and collaboration built-in.

Main Features

1. Shared & Validated Context: Hexis Cloud stores company knowledge in a structured, version-controlled wiki. Each page has a designated owner who must approve changes before they go live, ensuring all information is verified. This creates a reliable knowledge graph that both people and AI agents can query. It works by storing all content as commits in a Git repository controlled by the customer, providing a full audit trail.
2. Collaborative Skill Library: Instead of individual prompt engineering, teams can build, share, and iteratively improve reusable AI skills (like "write-a-proposal" or "analyze-RFI"). Anyone can suggest edits via a browser interface, but changes are only deployed after approval by the skill owner. This feature includes 58 ready-made skills to accelerate adoption.
3. Unified MCP Tool Gateway: Hexis Cloud acts as a central MCP (Model Context Protocol) server, connecting once to dozens of tools (HubSpot, Notion, Jira, GitHub, etc.) and exposing them securely to any compatible AI agent (Claude Desktop, Cursor, ChatGPT). Credentials are managed server-side, so agents never see sensitive keys, and tools are dynamically looked up only when needed to avoid context window overload.
4. Fine-Grained Access Control: Security is enforced at the file, folder, skill, and tool level. Administrators can set precise permissions (read, edit, no access) for individuals, teams, or roles (e.g., Sales can read pricing, Finance can edit it, Support has no access). This ensures AI agents operate strictly within the data boundaries of the human using them.

Problems Solved

1. Pain Point: Prevents "AI agent sprawl" and hallucinations caused by agents using outdated, incorrect, or unauthorized data. Eliminates the need for individuals to repeatedly connect and authenticate tools for each agent.
2. Target Audience: AI Transformation Leads, Engineering/Platform Teams managing AI tooling, IT/Security Teams requiring governance, and cross-functional teams (Sales, Ops, Support) looking to operationalize AI with shared processes.
3. Use Cases: Essential for scenarios like: a sales agent drafting a proposal using the latest GTM skill and approved pricing data; an ops agent answering site-specific queries using owner-verified context; a developer using Cursor to work on code with access to Jira tickets and GitHub PRs through a single, secure gateway.

Unique Advantages

1. Strengths & Limitations (Pros & Cons):

   • Pros: Exceptional governance model with owner-approval workflows and Git-based versioning. Drastically reduces the security risk of AI tool access via credential isolation. Promotes collaboration and consistency across teams. Transparent, open-source code (Apache 2.0) and ability to self-host.
   • Cons: Primarily benefits organizations already using or planning to use multiple AI agents heavily. The value is maximized with team-wide adoption; individual users may not need its full suite. Requires initial setup and definition of content ownership structures.

2. Key Alternatives & Differentiation:

   • vs. Individual MCP Setups: Without Hexis, each user configures their own MCP servers and tool connections, leading to inconsistency, security risks from scattered credentials, and no shared knowledge base. Hexis provides a company-wide, managed layer.
   • vs. Generic Internal Wikis (Confluence, Notion): While these store information, they lack native, secure integration with AI agents via MCP, structured approval workflows for context, and a centralized skill library. They are passive repositories, not active AI platforms.
   • vs. Other AI Agent Platforms: Many platforms focus on building custom agents. Hexis differentiates by focusing on the shared context, skills, and tools layer that any standard MCP-compatible agent (Claude, Cursor) can immediately use, avoiding vendor lock-in.

Frequently Asked Questions (FAQ)

1. What is Hexis Cloud and how does it work with AI agents? Hexis Cloud is a central platform that provides AI agents like Claude and ChatGPT with secure access to your company's verified knowledge, pre-built skills, and connected tools (like CRM or GitHub) through the standard Model Context Protocol (MCP), acting as a single, governed brain for all your agents.
2. How does Hexis Cloud ensure data security and governance? Hexis enforces fine-grained access control per file, stores all data in your own Git repository for full auditability, keeps API keys secure on its server (never exposed to agents), and requires owner approval for all changes to context and skills, ensuring strict governance.
3. Can I use Hexis Cloud with ChatGPT and Claude? Yes, Hexis Cloud is fully compatible with any AI agent that supports the Model Context Protocol (MCP), including Claude Desktop, ChatGPT via relevant plugins, Cursor IDE, and others, serving as a unified tool gateway for them all.
4. What is the pricing model for Hexis Cloud? Hexis Cloud offers a free tier for up to 3 users. For larger teams, the paid plan costs €10 per person per month (excluding VAT). The entire platform is also open-source, allowing organizations to self-host it for free.
5. How does Hexis handle changes to company knowledge or AI skills? Any team member can suggest edits to context pages or skills through the web interface. Changes are only published and made available to AI agents after being reviewed and approved by the designated owner of that content, maintaining a single, validated source of truth.