See everything
Agent activity, tool usage, data access, approvals, diffs, previews, and outcomes stay visible in one operating surface.
Tunc Hub is the enterprise platform for AI development, the heart of AI operations and agentic control.
Control what AI can do, and keep every model, skill, tool, approval, artifact, and chat tied to your operating model.
Agent activity, tool usage, data access, approvals, diffs, previews, and outcomes stay visible in one operating surface.
Govern models, tools, skills, personas, memory, shell behavior, and destinations by company and workspace policy.
Complete logs, retained chat history, release metadata, and policy versions make AI work explainable after the fact.
{
"allow": [
{ "provider": "openai", "models": ["gpt-5"] },
{ "provider": "anthropic", "models": ["claude-opus-4.8"] }
],
"tools": {
"shell.exec": "ask",
"git.*": "allow",
"external.push": "deny"
},
"storage": "core",
"retention": "company_policy"
}Tunc Hub gives leadership a control plane without taking the developer workbench away from the people building the software.
A single policy layer for models, tools, skills, data, personas, permission modes, and agent behavior.
Invite collaborators, share AI chats, assign work to agents, and keep decisions visible across the team.
Store chats, artifacts, releases, and searchable history in Core when the organization needs continuity.
Deploy Tunc Core in the cloud, private cloud, on-premises, or a local environment you control.
Use immutable policy versions, release records, approvals, denials, and retained history for operational review.
Keep the desktop fast while giving agents access to real repos, previews, tests, tools, and release flow.
Run Tunc Core as managed cloud, private cloud, on-premises, or local infrastructure. The same Hub desktop connects to the Core you choose.
Your domain, sign-in, storage, policy versions, retention, release channels, and audit windows stay under your control.
Start with the desktop. Add Core when the team needs shared control, retained history, local models, or private deployment.