Governance

Everything your agents do is authorized, logged and measured

Permissions by role, a history of every run, measured costs, and you decide where it all runs.

A structure that follows yours

From the organization to each person, every level sets its own rules.

Your organizationModelsToolsRoles
Sales space
Agents · 2
  • Lead qualification
  • Quote follow-up
People · 3
  • Julie M.Admin
  • Marc D.Builder
  • Sophie L.Viewer
Support space
Agents · 2
  • Ticket triage
  • Reply drafts
People · 3
  • Karim B.Admin
  • Léa P.Operator
  • Paul R.Viewer
Finance space
Agents · 2
  • Invoice follow-up
  • Payment matching
People · 3
  • Inès T.Admin
  • Nora S.Builder
  • Marc D.Viewer
  1. The organization

    The shared frame: who is a member, which models and tools are available, which roles exist.

  2. Spaces

    One space per team or department: its agents, its connections, its schedules and its log.

  3. People

    One role per space, from admin to viewer. You can also define your own roles.

Stay in control of every agent

Everything your agents do is authorized, logged and measured. And you decide where it all runs.

Role-based permissions

Per space, you decide who builds, who runs, who sees, and with which tools.

Full history

Every run keeps who started it, its model, its result and its cost. Permission changes are also recorded in the database; viewing them in the interface is on our roadmap.

Measured costs

The cost of every model call is counted and tied to the run and the agent. The total is shown on each run.

Any model

15 providers, commercial or local. No lock-in to a single one.

Your servers
AppstrateYour agentsLocal modelYour data
Outbound traffic under your control
Internet

Hosted on your premises

On your servers with a local model, even without Internet access. Only calls to the tools you plug in leave, plus reading the model catalog, which you can turn off.

Open code

Apache 2.0 licensed: your team or an auditor can read it all.

See all of governance

Work out the gain of each agent

Appstrate counts the runs and measures their cost. You add the time each one saves: the return is computed instead of guessed.

Ticket triage · sample calculation
1,240
runs
counted
6 min
saved per run
your estimate
124 h
given back to the team
calculated
$38
in models
measured

Frequently asked questions

Where is my data?
Where you install Appstrate: on your servers, even without Internet access, or on our cloud, hosted in Canada or France. What a model sees depends on the model you choose: a local model stays with you.
Who sees what the agents do?
Every run is logged and measured, and permissions are set by role. Admins and builders see who started what.
How does the team sign in?
With email and password, magic link, Google or GitHub. Enterprise SSO (SAML, Microsoft Entra) is coming soon, with the Enterprise license.
Are you SOC 2 or ISO 27001 certified?
Not yet: these processes have not started. The code is open, so your technical team can audit it line by line.

Let's see what it would look like on your side.

Try it on your own in a few minutes, or take thirty minutes with us to scope a first use case.

To try it

Try it for free

Create a cloud account and describe your first agent.

For a guided start

Let's talk about your case

Thirty minutes with us to scope a first use case and its gain.

To audit

Read the code

Apache 2.0. Your technical team can read it all, install it all, keep it all.