Features

What Ulmero does, today.

Every feature on this page is in the platform now. Certified partners build each customer's system from these features and from the modules.

AI agents

Agents that work inside the system, under your rules.

As many agents as you need

Each agent has its own name, persona and instructions. Instructions add up from the platform to your organization to the agent to the person, and no layer replaces another.

Your own AI provider

Connect any OpenAI-compatible AI provider, set up as named profiles, and change models when you like.

Knowledge bases

Each organization has its own knowledge base that its agents search for answers.

Approval before high-impact actions

Before a high-impact action, the agent saves a request with a preview and a cost estimate, and waits for a signed-in person with permission to approve it.

Memory between conversations

Agents keep a memory of past conversations and recall what is relevant.

Connections both ways

Agents use tools on outside servers through the Model Context Protocol (MCP). Outside agents work in the platform through its own MCP server, with only the permissions of their API key.

Email from agents

An agent emails a person or a list after a signed-in person approves, with an unsubscribe link in every message and a daily limit for each organization.

Agents that hand off work

Agents send messages to other agents, with limits that stop a loop.

Workflows

Multi-step AI workflows, started on a schedule, by an agent or through the API.

Voice

Push-to-talk speech to text, and replies read aloud.

Images and short videos

Agents make images and short videos from a prompt.

An agent on your website

Put a public agent on your website, send each page to the right agent, and hand a visitor from one agent to another with a summary.

Feedback and review

People rate replies and report problems. Reports go to a moderation queue, and your team reviews conversations and records how each one turned out.

Running the business

The tools the business runs on, in one system.

Accounts

Customer accounts, each with its own history of activity.

Planner

Projects, tasks, subtasks and goals with targets, for a person, a group or an account, with time tracked by timer or by hand.

Working hours and time off

Each person's working hours and time away, and a team page that shows them.

Company and people

Your organization's profile, branding, members and staff directory.

Custom fields

Add your own fields to people, groups and organizations without a code change.

Saved views

Save a list's sorts and filters as a named view, shown as a tab.

Dashboard

A home dashboard of widgets you arrange and resize.

Files

Files and folders, private or shared, stored on disk or in S3-compatible storage.

Datasets

Databases of your own, created new or adopted from an existing database, and queried with read-only SQL.

Data from your services

Bring data from a connected service into a dataset on a schedule, with a field map and a record of every run.

Google Analytics and Search Console

Agents read your Google Analytics 4 and Search Console data, read-only.

Notifications

In-app notifications with a live badge, and email and text messages (SMS) through the provider you choose.

Product tours

Guided tours of the app that you edit without a code change.

Subscriptions

Stripe subscriptions for each account.

Several organizations

Run several organizations on one installation, with a hard boundary between them.

Three languages

The admin app in English, Spanish and French.

Control and security

Every change recorded, every permission yours to set.

Permissions

Grant each feature to people and to groups. An agent acts only with the permissions of the person using it.

History and rollback

Every change is recorded in a tamper-evident, hash-chained history, and an administrator can roll a change back.

Audit logs

Sign-ins, failed sign-ins, password changes and access changes, and a record of every billable AI call.

API keys

Keys with their own permissions and an expiry date, or keys that act as one person.

AI guardrails

Prompt-injection and personal-data filters for AI calls through OpenRouter, and a password policy.

Rate limits

Limits on requests for the installation, each organization, each endpoint and each API key.

Consent

Versioned consent policies, a record of each person's consent, and guardian links for people who need a guardian's approval.

Data retention

Set how long each kind of data is kept. A daily job removes what is older.

Security settings

A security status panel, a list of allowed website origins that an admin edits, and password reset by email.

Operations

Run it yourself, or have a partner run it.

Backups

Backups of the whole installation to S3-compatible storage, and restore from a backup.

Scheduled jobs

Jobs that run on the schedule you set.

Settings without a restart

Many settings apply while the system runs, with no restart.

Health and logs

A standard health check address, structured logs, and a log viewer for Google Cloud Logging.

Updates underneath you

The platform improves and you pull the updates, while what was built on top keeps working.

See it on your own business.

Tell us about your business, and we match you with a certified partner who builds it on Ulmero.