What it does in your chat apps
Teams, Slack and Google Chat are triaged into one stream together with Outlook and Gmail. Each message is sorted by project, urgency, whether it needs a reply, and risk, and a “while you were away” digest catches you up. A reply is drafted in your voice and waits in a queue inside Triazz.
Chat is polled, with optional instant push for Teams, Slack and Google Chat. Attachments are not read.
Ask from the chat app you already use
From Slack, Teams or Google Chat you can ask the team bot, a knowledge bot about your repositories, or a colleague’s bot. For Slack, a Slack app of your own on one person’s bot answers plain commands such as digest, drafts, status and pause, and can save a reminder or prepare a Jira item for your tap. It has no model of its own, it answers only you, and it requires every request to be signed by Slack.
Tapping a draft’s card in Slack can send exactly the text shown, but only if you switch that on and confirm it.
Nothing sends without your say
Triazz starts in Shadow Mode, which reads and drafts only. Risky abilities such as sending chat or email are off by default. Per action you choose Do, Ask or Never, and one kill switch stops everything outward. By default every bot reply opens with a line saying it writes on your behalf.
Every step goes into a tamper-evident activity log. See the Security & Trust Centre for what it does and does not protect against.
One bot for each person, one for the team
Each person has a private bot and database, and the team has a shared bot. Each message goes to exactly one bot, with its own voice, mode and safety rules. A colleague’s bot can ask yours only if you turned sharing on and named who may ask. Mail, chats and notes live in a separate private database per person, so one person’s data cannot leak to another through a query mistake.
People sign in with their company Microsoft or Google account. An administrator registers the company’s Slack, Microsoft and Google apps once, so everyone just clicks Connect.
Where your chat goes
Triazz is self-hosted: your messages, the database and your connection credentials stay on the computer or server you run it on. What leaves is the specific text sent to the AI model you chose for a job and the calls to the services you connect. With a local Ollama model, no text goes to a model provider. See the Privacy page.
Honest limits
- Chat is polled unless you turn on instant push. Replies start as drafts inside Triazz and wait for your say.
- There is no mobile app, and a shared server is designed for roughly 50 to 100 people; larger organisations run one server per team.
- Triazz needs Node 22+ (and Ollama for local models), and is set up with us because the source is not public. We recommend starting with one pilot user.
- Email, tasks, browser QA and repository answers are part of the product too. See the email page and the knowledge bots page.
Want to try it with your team?
Write to us and we will talk it through.