Federis StudioAgent and workflow builder

Build governed agents on your own infrastructure.

Federis Studio gives enterprise teams a visual canvas, a Copilot that drafts workflows from a plain request, and a multi-agent runtime. Connected to Federis, every step runs inside your control plane.

Book a demo Compare the two editions
A workflow of five steps: a trigger, document retrieval, an agent, a human approval and an action. Policy, guardrail and audit checks apply at every step. Triggerschedule Retrieveyour docs Agentrunning Approvalhuman review Actiontool call Policy decision · guardrail check · audit event at every step Triggerschedule Retrieveyour docs Agentrunning Approvalhuman review Actiontool call Policy decision · guardrail check audit event at every step
A workflow on the Studio canvas. Policy, guardrails and audit apply to every step in both editions; on the platform they come from Federis.

Everything a team needs to build and run agents

Studio covers the whole path from a first draft to a monitored run, so builders and reviewers work in one place.

01

Visual canvas

Lay out agents, tools and decisions as a graph. Anyone on the team can read what a workflow does.

02

Copilot

Describe the outcome you want. The Copilot drafts a workflow on the canvas for you to review and refine.

03

Templates

Publish proven workflows to the whole organisation, or only to chosen departments, workspaces or data domains.

04

Multi-agent runtime

Run graph workflows with several cooperating agents. Runs are durable, so they can be resumed and replayed.

05

Integrations

Connect business apps and tools, including MCP servers, as steps in a workflow.

06

Approvals and run history

Pause a run for a human decision before a risky action, and review every run afterwards.

One Product, two editions

Run Studio as its own product, or as a module of Federis that uses the identity, models, policy and audit you already govern.

  Standalone Platform, with Federis
Runs as Its own product An optional module of Federis
Sign-in Google, email or direct sign-in Your Federis identity provider (OIDC)
Models Providers you connect Only models deployed on Federis. No provider keys, no fallback model
Policy, guardrails and retrieval Studio's own Federis policy, guardrails and retrieval
Audit Studio's own ledger Kept in Studio and sent to the Federis audit vault
Risky actions Studio approval steps Governed Actions: request, decision, then execute

Every workflow is governed by default, in either edition.

Policy

Policy and guardrails on each step

Every step is checked before and after it runs. Standalone uses Studio's own policy engine and guardrails; on the platform the decisions come from Federis.

Approvals

Humans approve risky actions

Pause a run for a decision before a risky tool call. Standalone uses Studio approval steps; on the platform it runs as a Governed Action: request, decision, then execute.

Audit

Audit that holds up

Every run is recorded in Studio's own ledger. On the platform the events also reach the Federis audit vault.

Models

Only approved models

On the platform, pickers list only the models deployed for the user. If a step's model is not allowed, the run stops before it starts and names the model.

Bring Studio to your control plane

We will walk through Federis and Federis Studio against your environment and your first workflow.

  • Visual canvas and Copilot
  • Human approvals
  • Durable runs and replay
Book a demo

Or see the services