AI employees for the way software teams actually ship.
Etenos builds role-specific AI employees that understand your context, work through real workflows, and operate under explicit policy — observable, approval-gated, and auditable in production.
Not chatbots. Employees.
Each Etenos product is a role-specific AI employee — it understands company context, works through real workflows, integrates with your tools, and operates under explicit policy. Syntaxis is available today; the rest of the roster is in active development.
Etenos Syntaxis
A desktop AI engineer that plans work, prepares diffs, validates, and opens PRs — pausing before every side effect.
Etenos QA
Regression planning, test-case generation, bug reproduction, release validation, and quality reports.
Etenos Analyst
Requirements gathering, ticket refinement, process mapping, acceptance criteria, and stakeholder summaries.
Etenos PM
Roadmap synthesis, prioritization, specs, sprint planning, stakeholder updates, and release coordination.
One runtime. Every role.
Etenos is a role-based AI employee runtime: a local agent that does the work, and a cloud control plane that handles accounts, organizations, policy, integrations, and audit — without raw code ever leaving the machine.
Every side effect stops and asks.
Read-only by default. Branch, patch, validation shell, commit, push, and pull request each queue to a single approval inbox. The pause is the product.
Every side effect waits here · local writes, git ops, GitHub mutations.
Built for real production teams.
Observable
Every run is a legible sequence of stages and approvals — not an opaque agent that disappears and returns with a surprise.
/01Approval-gated
Nothing writes files, runs shell commands, commits, pushes, or opens PRs without an explicit, per-action approval.
/02Auditable
Approvals, executions, policy changes, and trust changes are recorded in a local audit log you can review and export.
/03Workflow-native
Maps onto your actual process — GitHub Projects, status fields, assignees, branch rules — instead of a parallel dashboard.
/04Local-first
Source code and command output stay on the developer's machine by default; the control plane sees sanitized summaries.
/05Policy-driven
Strict, standard, or trusted policy profiles decide which side effects need approval and which can run automatically.
/06Start with Syntaxis.
The AI Software Engineer that plans, drafts, validates, and opens PRs — pausing before every side effect. The rest of the roster is on the way.