An AI that acts
across your whole suite.
Ulti is your AI. It sees what you give it, and runs on the model you want.
The context is already there
An in-app chat that knows the open mail, selected file or current event, in a side panel or full screen.
Everywhere in the suite
Side panel in Ultimail and UltiDrive, or full screen.
Suite context
Understands the open mail, selected file or current event.
Visible sources
Each answer shows the mail and files it used and the tools it called.
History and conversations
Resume threads, pin useful ones, start again from a prompt.
It can also act
Via MCP, UltiAI acts on your mail, files, calendar and contacts, with permissions per tool group.
Mail tools
Search, read, send, label and delete messages.
Drive and Contacts tools
Files, folders, shares, address books and cards.
Calendar tools
UltiCal events, invites and UltiMeet video links.
Suite and web search
Mail, drive and contacts index. Web search via self-hosted SearXNG or Brave Search: only the query leaves, never your mail.
Rule-based triage, model of your choice
Wire AI into your mail rules: LLM filing, summaries, reply drafts and API tokens for your agents.
LLM triage per rule
Custom prompt and context per rule; the LLM picks the label.
Summaries and replies
Long-thread recap and reply drafts in one click.
API tokens with partial permissions
For external agents: read some folders, send, file.
Provider per rule
Pick the model that fits each automation.
Your models, your choice
UltiAI speaks the OpenAI-compatible API: any compatible provider or server works. Switch models without migrating your data.
OpenAI
GPT and compatible APIs
Personal
Enter your OpenAI API key and pick the model for chat and triage. Requests then go to OpenAI.
Enterprise
Azure OpenAI or an enterprise gateway on the same compatible endpoint.
Mistral AI
European models
Personal
Mistral models hosted in Europe, set up in a few seconds.
Enterprise
Dedicated Mistral deploy or on your server via a compatible endpoint.
Anthropic
Claude family
Personal
Connect Claude through an OpenAI-compatible gateway. Requests then go to Anthropic.
Enterprise
Org keys and quotas managed in admin.
Ollama
Local and self-hosted models
Personal
Run Llama, Qwen or Mistral locally: inference stays on your machine.
Enterprise
Internal inference server (vLLM, Ollama) shared across the organization.
OpenAI-compatible API
Any conforming endpoint works: cloud, gateway or local server.
Local and self-hosted
Keep inference on your side when data must not leave.
Permissions per tool
Turn each MCP tool group on or off for the assistant.
Switch freely
Move from one provider to another without migrating your data.
Connected to suite apps
UltiAI integrates natively with other UltiSuite apps, same identity, contacts and storage.
Try it free, 10 GB included
All 10 apps are in every plan. Export your data if it is not a fit.