Why Tempa
One control plane for every coding agent
Your developers use Claude Code, Codex, GitHub Copilot, Gemini CLI and Antigravity, often side by side. Tempa gives each of them the same standards, guardrails and company context, per team, from one place, on the subscriptions they already have.
Every agent has its own settings
Each coding agent has its own instructions file, its own skills format, its own hooks and its own MCP configuration. A standard written for one agent does not reach the next, and a guardrail set in one vendor's admin console stops at that vendor.
Teams differ too. The payments team needs stricter rules than the prototyping team, and a client project needs its own context. One configuration for the whole company does not fit either.
Describe it once, deliver it everywhere
One profile per team
Standards, skills, commands, subagents, policies and MCP endpoints per organisation, department, team and project. Each level inherits from the one above, and modes say whether a team may add to, change or only tighten what it inherits.
Each harness in its own format
Tempa renders the profile for Claude Code, Codex, GitHub Copilot, Gemini CLI and Antigravity, and says per harness what it enforces natively and what it approximates.
Decisions on every tool call
The same policy bundles decide for every harness: allow, ask or deny, with a reason and what to do instead.
One view of the result
Adoption, policy decisions and delivered work per team, across harnesses, and an audit log of every change.