Git-backed workspace
Hexis keeps skills, tool manuals, and knowledge in a git repository you control, so the workspace can be reviewed and changed using normal branch and change-request workflows.
Hexis
Hexis is an open-source, git-backed workspace for AI agent skills, tools, and knowledge. Govern access, review changes, and connect MCP-capable agents from one repo.
What Hexis is
Hexis is an open-source workspace for AI agents that keeps a company's skills, tools, and knowledge in one git-backed place. According to the README, it is the open-source core of the Bevel platform and is meant to be centrally managed, reviewed, and access-controlled while still being usable from any AI agent.
The product is built for teams that want agents to work from the same governed knowledge base. It supports a built-in remote MCP server, role-based access, change requests with owner approval, and deployment to an empty git repository on any host, including GitHub, GitLab, Bitbucket, Azure DevOps, or self-hosted git services.
Key capabilities
Single-agent connection
The product connects to MCP-capable agents through a single connection key and a built-in remote MCP server, so agents can use the workspace without per-tool credentials or separate setup projects.
Review and access control
Changes are reviewed before they land. The README describes change requests, owner approval, and role-based access so agent updates follow the same control model as human edits.
Tools with secret storage
Tool manuals use UTCP and an encrypted secrets vault, giving the workspace a way to store tool instructions alongside protected secrets.
Vendor-neutral agent support
Hexis is independent of one agent vendor and is meant to work with Claude Code, ChatGPT, Cursor, or other MCP-capable agents, which helps teams switch or mix agents by task.
Bootstrap from an empty repository
The repository is seeded from a starter template on first run, and the setup flow can initialize an empty repo and write a roles file for the first admin.
Common ways teams use Hexis
Centralize governed agent knowledge
A team keeps its agent-facing playbooks, tool instructions, and knowledge in one reviewed repository instead of scattering them across separate systems.
Connect agents without per-tool setup
An organization gives employees a single connection key so their chosen agent can access only the skills and tools allowed by their role.
Review updates before rollout
When a skill or tool manual needs to change, contributors submit a change request and owners review it before the update lands on protected branches.
Evaluate the workflow before self-hosting
A company tests the public demo to see the workflow end to end before deploying its own instance or asking for a managed deployment.
Mix agent vendors by task
A team chooses one agent for some jobs and another for different tasks while keeping the underlying workspace and repository the same.
Pros and Cons
Pros
- Keeps skills, tools, and knowledge in one git repository that the company owns.
- Supports review and approval workflows before changes land, which can help teams govern agent-facing content.
- Works with multiple agent products rather than tying the workspace to one vendor.
- Can be deployed with Docker Compose and initialized from an empty repository.
- Includes a public demo that users can try before self-hosting.
Cons
- The source does not show detailed integration documentation beyond MCP-capable agents and the named examples in the README.
- Managed hosting is mentioned, but pricing and service terms are not listed on the page source.
FAQ
Which AI agents can use Hexis?
Hexis is designed to connect to Claude Code, ChatGPT, Cursor, or any MCP-capable agent through a single connection key. The README describes a built-in remote MCP server using OAuth 2.1.
How is Hexis set up?
The README says the app can be deployed in about 5 minutes with Docker and Compose. It needs an empty git repository, an admin email and password, and two generated secrets; public deployments also need backend and frontend origin URLs.
How does Hexis manage knowledge and changes?
Hexis stores skills, tool manuals, and knowledge in a git repository you own. It uses branches, change requests with owner approval, role-based access, and an encrypted secrets vault for tool manuals.
Is there a demo or trial available?
The project README describes a live demo at demo.bevel.software. It is a public instance you can sign into with a Google account and is populated with fictional company knowledge, skills, and tools.
Does Hexis offer a managed deployment option?
The README notes that a managed instance is available from the Bevel team, including hosting, upgrades, backups, and SSO. Pricing is not listed in the source.
Quick Facts
- Category
- Developer Tool
- Source domain
- github.com
- License
- Open source project
- Deployment
- Docker Compose; self-hosted or managed
- Primary users
- Teams managing AI agent skills, tools, and knowledge
- Agent protocol
- MCP
Hexis Alternatives
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads in Firecracker micro-VMs with private networking and SDK, CLI, or MCP control.
AakarDev AI
AakarDev AI helps teams manage AI provider access, project setup, logs, and analytics in one dashboard. BYOK support included.
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repos and verifies changes with compilers, debuggers, sanitizers, and tests.
Goldfish
Goldfish is an AI memory app for macOS that helps you reply, write, summarize, and continue with context already on your Mac. It stores memory locally and writes in your tone.
Lasso
Lasso is an ecommerce product data platform for enriching catalog records, processing supplier files, generating product content, and tracking competitors. It includes a web app, REST API, SDK, and MCP server.
ArtDeck
ArtDeck is a visual reference board app for iPhone, iPad, and Mac. Collect mixed references on one canvas and study them with ToolBox lenses. Offline, optional iCloud sync, one-time purchase.