Give each AI persona its own name, prompt, knowledge base scope and set of tools, chat with them directly, and trigger them from Telegram commands that run your automation flows.
Configure once, then talk to it or let it respond automatically.
Name it, give it an avatar, a system prompt and a model — each agent is its own configuration, not a shared one.
Restrict a persona to specific knowledge base categories, so it only answers from what it should know.
Attach specific tools (backed by RPC functions) a persona is allowed to call — some can require confirmation before running.
Talk to any active persona directly, or wire a Telegram command to it so it responds automatically and can kick off a flow.
Chat tab, Agents tab, Activity tab — one page, one team.
Talk to any active persona in a chat interface — useful for testing a prompt or just getting an answer.
Create, edit and deactivate personas — each with its own model, language, memory length and confidence threshold.
See what your agents have actually done, not just what they were asked to do.
Bind a persona to a Telegram command, bot message or inline button — it can hand off straight to an Automation Hub flow.
Set how many turns of context a persona keeps, so multi-step conversations don’t start from zero every message.
Optionally inject live workspace data into the prompt, so answers reflect your actual business, not a generic template.
The same system, configured for different jobs.
Scoped to your knowledge base, answering questions the way a specific team member would.
Responds to a bot command and can launch an automation flow instead of just replying with text.
Given the right tools, it can look up or act on real workspace data when asked, with confirmation gates on the sensitive ones.
Other Oprela modules that pair with this one — same subscription, same data.
AI Agent Team is part of Automation Hub in Oprela — credit-metered.