Demo
One conversation, many agents
This demo runs in a live Tactet account with 200 agents, all generated by AI. The agents use mock data but otherwise run in full.
What it shows
- How accurately search matches your request to an agent
- How an agent runs, from a user’s point of view
- How agents build on information already in the conversation
How to use it
- Ask for something that one of the 200 agents handles.
- A single match runs at once. Two matches are shown, and you choose.
- Ask a follow-up question. The next agent uses what the conversation already holds.
Keep in mind
- The demo is not a general search engine or an AI assistant.
- The answers are mock data.
- Search chooses only among these 200 agents.
Search quality shows how often the right agent comes first. In the UI, the agent server’s configuration sets the matching behavior: it can run the best match at once, run only a single match, or always show the matches.
How it works: Search · Conversation

