Question 1

Is QubStack a closed no-code builder?

Accepted Answer

No. QubStack generates real source code and exposes the project through Git repositories, Docker files, configuration, logs and deployment controls.

Question 2

Can I build my own AI agents on QubStack?

Accepted Answer

Yes, and in two ways. Development agents build and maintain the project: a master agent holds the architecture and delegates to specialised code, git, infrastructure and chat agents, all visible on a graph where you can add, remove, connect them and edit their prompts. Service agents are agents your own application calls as OpenAI-compatible endpoints — you write the prompt, pin the exact model, and see the cost, call count and recent exchanges of each one, billed to that project.

Question 3

Can I choose which AI model an agent runs on?

Accepted Answer

Yes. Each service agent is pinned to a model you choose — a specific model or a capability class. If that model is unavailable the call fails plainly rather than being answered by a different model, so the answer you get always comes from the model you asked for.

Question 4

Can I deploy outside QubStack?

Accepted Answer

Yes. Projects can run on managed containers, your own VPS or a machine at home, with SSH host registration and deployment automation.

Question 5

What does QubStack automate?

Accepted Answer

QubStack automates architecture planning, code generation, AI review, builds, deploys, HTTPS routing, domains, logs, monitoring and self-healing fixes.

Question 6

Does QubStack support AI agents and MCP?

Accepted Answer

Yes. Projects can use Chat Agent, project agents, multi-agent workflows, A2A links and MCP context so agents understand the platform, project files and connected services.

Contact us