Git is still the center of gravity
No matter how much agent tooling changes, GitHub remains the canonical source of code, history, and collaboration. The fastest path to adoption is not inventing a new workflow. It is turning an existing repository into a runtime with as little ceremony as possible.
That is why Tennant is treating GitHub deploys as a core primitive instead of a bolt-on integration. If a team can already point to a repo and a branch, they should be able to provision a sandbox and get to usable output quickly.
Deploy should lead straight into inspection
A deployment is not the finish line for agent infrastructure. The next step is almost always inspection: read logs, execute a follow-up command, compare outputs, or issue another run against the same environment.
We want Tennant deploy flows to collapse those steps into one surface. Queue the deployment, watch status progress, open the runtime, and continue from there through the same sandbox record.
Why this matters for AI teams
Agent workflows tend to loop between code changes and runtime validation far more frequently than standard web deploys. The lower the friction between repo state and runtime state, the more useful the platform becomes for evaluations, coding agents, internal tools, and production automations.
That is also why the dashboard, HTTP API, CLI, and SDKs all need to speak the same lifecycle. GitHub deploy is not just a button. It is an addressable action across the entire control plane.