Platform · Works with your AI
Works with the AI you already use.
Claude, ChatGPT, Microsoft Copilot Studio, developer tools, and the agents you build, all using the same actions, which stay when you change assistants.
Assistants
Build once. Use it in every assistant.
Your people keep the assistant they already use. Add Bazinga to each one as a connector, and the same tools and rules are there in all of them.
Claude
A connector, for one person or everyone.
Add Bazinga as a custom connector by its address. On Team and Enterprise plans, an owner adds it once for the whole organization. Works on the web, desktop, and mobile.
Sign-in: Each person signs in as themselves.
ChatGPT
A custom MCP server.
Add Bazinga to ChatGPT as a custom MCP server, on the ChatGPT plans that support one, within the policies your workspace administrators set.
Sign-in: Each person signs in as themselves.
Microsoft Copilot Studio
Part of the agents you build.
Add Bazinga to a Copilot Studio agent as an MCP server, and the agent gets your actions.
Sign-in: Personal sign-in, or a service key for the agent.
Developer tools
Claude Code, Cursor, and others.
Configure Bazinga once in the tool. When it runs on the developer’s own machine, it can reach Bazinga inside your network too.
Sign-in: Personal sign-in, or a key.
Your own agents
Agents built with an SDK, n8n, and scheduled jobs.
Any agent or tool that can connect to a remote MCP server connects the same way, with a key of its own.
Sign-in: A service key, running as the account you chose.
One toolkit
Different teams, different assistants. One set of rules.
Finance prefers Claude, sales lives in ChatGPT, and IT builds agents in Copilot Studio. They all reach the same systems the same way, and each person still gets only what their role allows: the finance manager sees every invoice, the sales rep none.
The same actions everywhere.
Build an action once, and use it in every assistant you connect. Each assistant’s own settings still decide which tools its people see.
The same rules everywhere.
Permissions, limits, and the record don’t depend on which assistant a person prefers.
The same record everywhere.
One audit trail, with the assistant named on every line.
Standard
Built on the Model Context Protocol.
Bazinga speaks the protocol assistants use to connect to tools, so you’re never tied to a single AI provider’s roadmap.
An open standard.
The Model Context Protocol is the common way assistants connect to tools and data, supported by the major AI providers.
Built apart from the protocol.
The protocol keeps evolving. Your actions are built in Bazinga, not in it, so they stay the same when it changes.
No assistant lock-in.
Change assistants, or use several at once, without rebuilding a single action.
Reach
Cloud assistants. Private systems.
Where the assistant runs decides how it reaches Bazinga. Each case has a path that keeps your systems private.
Cloud assistants.
Claude, ChatGPT, and Copilot Studio call Bazinga from their providers’ clouds, so your Bazinga server answers on the internet, behind your rules.
Systems inside your network.
Bazinga can run on your own servers, beside the systems it connects. Cloud assistants then need a secure way in to Bazinga alone, and the systems behind it stay off the internet.
Developer tools.
When Claude Code, Cursor, or a similar tool runs on a person’s own machine, it reaches Bazinga over your internal network or VPN, with nothing opened to the internet.
Related capabilities
Bring the assistant your teams already use.We’ll connect it to your systems.
Claude, ChatGPT, Copilot Studio, or your own agents: see Bazinga working in it, with your rules.