About VAKEVA

VAKEVA lets a team compose AI agents from a role and a model, connect them to the data they need, and run them as scheduled workflows — without operating any of the infrastructure underneath.

Every agent is a pairing of a role — a description, an instruction set and a set of skills, in the same shape our own working team already uses — and a model best suited to that role, suggested automatically and always visible to you.

Models are open-weight: Qwen, Mistral, and GLM*. The infrastructure they run on, and what it costs us to run them, is ours to manage and never shown to you — you see your agents, your workflows, and your data.

* GLM models are coming soon.

Every access your agents make to your data is recorded and shown back to you in plain language, tied to the exact agent, workflow and right that made it.

VAKEVA is pre-launch. See the Terms of Service and Privacy Policy for what that means today.