Step 1
Keep it in sqemes
Write skills, prompts and personas once. Every saved change keeps a version, and you decide who may see and who may edit each one.
Claude is where many teams write, analyse and code. sqemes keeps your team's skills, prompts and personas once and brings them into Claude, Claude Code and every other AI tool you use.
Claude can connect remote MCP servers as custom connectors on every plan. That is how sqemes works inside Claude. The browser extension works on the Claude web app as well.
Three steps on Claude Team or Enterprise.
Step 1
Write skills, prompts and personas once. Every saved change keeps a version, and you decide who may see and who may edit each one.
Step 2
On Team, an owner adds sqemes as a custom connector. Then each member connects their own account. On Pro and Max, everyone adds it themselves.
Step 3
In Claude, Claude Code and every other tool connected to sqemes, the AI loads the playbook the task needs. A change in sqemes reaches all of them.
| Feature | sqemes | Claude |
|---|---|---|
| What it is | A library for AI know-how: skills, prompts and personas, used in other AI tools and in sqemes Chat | An AI assistant with chat, projects, skills and Claude Code |
| Where the know-how is used | In Claude, ChatGPT and Gemini through the browser extension, any MCP client, sqemes Chat | In the Claude apps and Claude Code, and in Claude's Microsoft 365 add-ins |
| Reusable knowledge | Skills with files, loaded when a task needs them | Skills that load when a task calls for them; projects with their own knowledge |
| Sharing | Per playbook: everyone, named people and groups, or only me; who may edit is set separately | Owners provide skills for everyone; projects are shared with view or edit rights on Team and Enterprise |
| MCP | An MCP server for your playbooks and personas; sqemes Chat also connects other MCP servers | Connects remote MCP servers as custom connectors on every plan; on Team, owners add them |
| Models | Any provider with your own key, or the plan's AI credits | Anthropic's Claude models |
| Training | Your content is never used to train a model | Inputs and outputs of the commercial products are not used for training by default |
| Hosting | Hosted in the EU; self-host on your own servers under a fair-code license | Data stored in the US, according to Anthropic |
As of: 15 November 2026, from Anthropic's own pages. Sources: Anthropic: custom connectors, Anthropic: skills, Anthropic: projects, Anthropic: server locations, Anthropic: model training.
Most teams use both. This helps if you wonder whether Claude alone is enough.
your whole team works in Claude and nowhere else.
your know-how should work in Claude and beyond.
What teams ask about sqemes and Claude.