Agent Serve · Enterprise
Agent Serve is invite-only. Your teams build and run AI agents behind your SSO, in your own cloud if you choose.
Agent Serve is the AI workspace where enterprise teams build, deploy, and manage AI agents. Accounts are created by invitation. Organizations sign in with SAML or OIDC single sign-on, can self-host in their own cloud, run models with zero data retention through supported providers, set roles and access policies per team space, and trace every agent run.
The foundation
Agent Serve gives security and IT teams single sign-on, self-hosting and zero data retention before the first agent runs.
Agent Serve supports SAML and OIDC single sign-on, with SCIM 2.0 to provision people and groups automatically.
Deploy Agent Serve inside your own cloud account or VPC, next to your data, with the same product your team sees here.
Run models under your organization’s zero data retention terms, or keep inference inside your network, and decide how long Agent Serve keeps logs.
Zero data retention applies with supported model providers. Some provider features keep data by design, such as Gemini Deep Research, so they are outside zero data retention.
Getting started
Agent Serve accounts are created by invitation, so every organization starts with a conversation, not a sign-up form.
Tell us what you want to build. A person reads every request and replies.
We talk through the agents you want and the data they need.
We create your organization and invite your first admins.
Your admins connect SSO and invite the rest of the team from inside Agent Serve.
In depth
Agent Serve puts controls where agents are built, deployed and run, so governance is part of the work rather than a review at the end.
Agent Serve gives each team space admin, write and read roles, and access policies limit which blocks, models and integrations a group can use.
Agent Serve Relays ask before they send email or book meetings, and a workflow can pause for a person at any step.
Agent Serve traces each workflow run block by block, with inputs, outputs, tokens, cost and errors.
Agent Serve ships a Helm chart, Docker images and reference Terraform for AWS, so you can run it in your own cloud.
For the whole organization
An Agent Serve organization holds many team spaces, each with its own members, data and agents.
Credit risk · Claims · IT
Organization admins manage the usage limits of each Agent Serve team space as more people build agents.
Settings → Usage
Agent Serve uses your organization’s own model provider keys, set at organization or team space level.
Settings → Model API keys
Agent Serve keeps data from one day to five years, or forever, per organization or team space.
Log retention · 1 year
Agent Serve redacts personal data in logs, workflow input and block outputs before it is stored.
Logs · Workflow input · Block outputs
The Agent Serve Activity audit lists organization and workspace changes and exports them as CSV.
audit-log.csv
Before and after
| Work | Before | With Agent Serve |
|---|---|---|
| Who uses AI | BeforePersonal accounts on consumer AI tools, outside IT’s view. | With Agent ServeInvited people, signed in through your SSO, in one Agent Serve organization. |
| Where data goes | BeforePrompts sent to whatever model a person picked. | With Agent ServeYour provider keys, zero data retention with supported providers, or models in your network. |
| What agents may do | BeforeAn assistant acts with whatever access its key has. | With Agent ServeAccess policies, row-level security and approvals before external actions. |
| Evidence for audit | BeforeScreenshots collected by hand before each review. | With Agent ServeAn exportable activity audit and a trace of every run. |
Under the hood
| Access | By invitation. Request access, then admins invite the team. |
|---|---|
| Single sign-on | SAML 2.0 and OIDC with verified domains. SCIM 2.0 provisioning. |
| Sessions | Maximum session lifetime, idle timeout, and Sign out all members. |
| Roles | Admin, write and read per team space, plus permission groups. |
| Hosting | Agent Serve cloud, or self-hosted with Helm, Docker or Terraform. |
| Models | Your own provider keys, or self-hosted models through Ollama, vLLM or LiteLLM. |
| Data retention | 1 day to 5 years, or forever, per organization or team space. |
| Compliance | SOC 2 ready and GDPR ready. Not a certification. |
Governance
Each Agent Serve control below is a setting in the product today, worded the same way on every page.
Security and governanceSign in through your identity provider with SAML 2.0 or OIDC, and provision people with SCIM.
Admin, write and read roles per workspace, plus permission groups that limit which models, blocks and integrations a team can use.
Relays never send email or book meetings without an approval, under any autonomy setting. An approval covers the exact payload the reviewer saw and expires if nobody decides.
Proposals, approvals, declines and executed actions are written to the workspace audit log, which admins can export.
Zero data retention with supported model providers, backed by your own provider keys and agreements or self-hosted models. Some provider features keep data by design, such as Gemini Deep Research, so they are outside zero data retention.
Run Agent Serve in your own cloud or data centre, next to the systems it reads.
The controls SOC 2 looks for are in place. Agent Serve does not claim a SOC 2 report or certification.
Retention from one day to five years, deletion and access controls support GDPR obligations. Ready means the controls are in place, not a certification.
What teams build
AI coworkers that ask before they act outside Agent Serve.
Agents built visually, deployed as APIs, chat pages or tools.
Build and manage agents in natural language.
A block-by-block trace of every run.
The full list of security and compliance controls.
Further reading
Disputes that fail reconciliation tend to sit in a side queue until the chargeback window has closed. Here is an end-to-end setup that finds them early, ranks them by deadline and drafts the pack, with an analyst signing every filing.
Cards and payments operations · 12 min readPeriodic reviews pile up because every file starts from scratch, even when nothing has changed. Here is a setup that refreshes the data, drafts the file and routes it by risk, so analysts spend their time on the reviews that need judgement.
Retail and business banking · 12 min readSampled call reviews find a missing risk warning weeks after the client acted on the advice. Here is a setup that checks every recorded advice call within a day and puts the quote in front of a compliance reviewer.
Wealth and advice · 12 min readQuestions
Agent Serve accounts are created by invitation. Request access, and we set up your organization and invite your first admins, who invite the rest of the team.
Yes. Agent Serve ships a Helm chart, Docker images with a production compose file, and reference Terraform for AWS.
No. Agent Serve is SOC 2 ready and GDPR ready: the controls are in place, but Agent Serve does not hold a certification.
Agent Serve offers zero data retention with supported model providers, backed by your own keys and agreements or self-hosted models. Some provider features, such as Gemini Deep Research, keep data by design.
Last updated
Enterprise access
Agent Serve is available to teams by invitation. Tell us what you want to build and we will set up your workspace with you.