Skip to main content
The Chat Window is Enconvo’s main space for ongoing conversations. Choose an Agent, add the material it needs, and keep related work together.

​
Start a conversation

  1. Open Chat Window from Enconvo.
  2. Select an Agent and create a New chat.
  3. Type your request and attach any relevant files, app context, or knowledge-base material.
  4. Check the model and reasoning selection.
  5. Send the request and follow tool progress if the Agent is using tools.
Use a complete request: “Summarize the attached proposal in five bullets and list open questions” gives the Agent both input and an expected result.

​
Organize work

Use the sidebar to return to recent conversations and organize related conversations in projects where available. Check a conversation’s menu for renaming, moving, archiving, or deletion actions supported by your build. Read the confirmation before deleting history.

​
Model choices

A top-level New chat can set the Agent’s model default. A new conversation inside a project uses that project’s model preference. Each created chat can keep its own model and reasoning override. See Model Selection before changing defaults used elsewhere.

​
Context and attachments

Use only the sources relevant to the request. A file attachment, screen capture, knowledge-base lookup, and browser page provide different kinds of context. See Context Awareness for source selection and Files and Folders for Agent file access. The Agent’s Memory and a Knowledge Base have different purposes: memory holds reusable facts; a knowledge base stores material for retrieval.

​
Read and reuse responses

Expand work details to inspect tool calls and their results. The reply action row offers Copy, Read Aloud, and More in current builds. Visual Answers can embed diagrams or interactive content directly in a reply. Use Dictation to compose with speech, or Live Voice for a two-way spoken conversation.

​
Troubleshooting

If a response uses the wrong model, inspect this conversation’s selection as well as the Agent default. If the Agent cannot see a file or screen, review its attachments and permissions. If a tool is unavailable, check the Agent’s enabled tools instead of repeatedly rephrasing the same request.