A CRM with an AI agent inside: customers, leads, deals, invoices and service in one place — and it is the agent that keeps the data, not you
This is a customer database (CRM) on your own server where an AI agent is a full participant, not a bolt-on: it creates and cleans customer records (contacts, companies, registration details), keeps the communication timeline, merges duplicates, accepts requests from your website, moves deals through the pipeline, prepares quotes and invoices, runs mailings and communications, and services tickets on schedule. The key feature of the model: there is not a single editing form in the interface — you watch through the screens and give commands to your AI agent (a chat with Agent), while changes are made through the management key over the API (the management address your agent uses). The second feature: the product is assembled from a small core and add-ons you enable as you need them. Everything lives on your servers, and the data never leaves them. And when there is no work to do, the product spends neither processor time nor memory: it has no timer-driven background work.
The AI agent keeps the data: it creates and cleans records while you watch and give commands
Not a single editing form in the interface — you cannot spoil the data by accident
One customer base: contacts, companies, registration details, links and a holding hierarchy
Every record is checked for duplicates, with a separate queue for reviewing them
Duplicates merge with a rollback, and the AI brings the data to one consistent form
The whole history in one feed: notes, calls, emails, tasks and files
CSV import and export with a dry run, rollback and a history of operations
One search field finds customers, deals and tickets alike
Request intake in one place: public addresses by key and a ready-made form page
No lead is left without an owner: distribution rules and first-touch deadlines
Pipelines and stages are configured as data, and every move goes into a log
From quote to cash: goods, discounts, invoices, payments, receivables, contracts and subscriptions
Mailings and campaigns with consents, audiences and honest statistics
Email, calls, messengers and a website widget in one conversation with a shared inbox
Service tickets with deadlines, escalation, a knowledge base and a customer portal
Reports and dashboards are set up as data, exported to tables and delivered by email
The product is built from a core plus add-ons: enable only what you need
Custom fields, custom objects for your industry, tags, views and segments
Observation screens update in real time and use words instead of codes
One management key for the agent: the browser can only read, only the agent can change
A self-documenting reference: an agent with an address and a key acts on its own
A chat with the AI agent inside the CRM: manage data with words without leaving the window
Automation: when-this-then-that rules and multi-step processes with a journal and approvals
Your own Python code when the built-in features are not enough
Everything on your server, and the license limits only the chat with the agent
Here is the whole product. Click any add-on to open its page: why you need it, what it does, what it is connected to and what it deliberately does not do.
The product is built from a small core plus twenty-three add-ons: sales, money, marketing, communications, service, plus automation, analytics and your own scripts
A fresh install starts with the core, and you enable the parts you need in the Add-ons catalog
Disabling keeps the data: the code is switched off while everything you built stays in place
Enabling and disabling take effect on restart, which is safer than changing the set on the fly
A 360° customer card with tabs: timeline, related records, files, notes, fields and change history
Not a single data editing form: in the browser a person only approves approvals and tasks, the records are changed by the agent, and everything updates in real time
