Product feature

Conversational agent

The conversational agent creates, edits, and synthesizes files in the context repository from a chat prompt.

What is a conversational agent?

The conversational agent is Qontext’s AI agent for creating, editing, and managing files inside the context repository through a chat interface. From the Context tab, a user opens the chat, describes what they want, and the agent creates folders, writes new files, edits existing ones, moves files, or deletes what’s no longer needed. It can also browse and search the repository by meaning or exact text, and combine repository content with synced source data to write consolidated or aggregated documents.

Why use a conversational agent?

Keeping a context repository organized and current is ongoing work: files need restructuring as a workspace grows, and content drawn from synced sources needs to be compared against what’s already written and refreshed when it’s gone out of date. Doing that by hand means someone has to notice what changed, find every file it affects, and edit each one manually. The conversational agent does that work from a single prompt, reading across the repository and its synced sources to make the edit or write the document directly.

How does the conversational agent work?

The conversational agent operates with the permissions of the person using it and never beyond them, so the same prompt can return different results for different users depending on what each one can access. To use it, open the Context tab, click the chat icon in the top right corner, and type the request.
Every edit the agent makes shows up in the Changes tab with a diff and version history, the same as an edit a person made directly. Files marked “Require review” hold the agent’s changes pending approval before they apply; everything else applies automatically. From the chat dropdown, a user can stop an in-flight response, start a new chat, or resume a previous conversation. Because the agent reads and writes directly against the context repository, every change it makes becomes part of its version history.

FAQ

Can the conversational agent see files I don't have access to?

No. The agent works with your permissions and never beyond them, so a file you can't reach won't appear in its results either.

Do the conversational agent's edits need approval before they apply?

Only on files marked "Require review." Those hold the agent's change pending approval in the Changes tab; everything else applies automatically.

Can the conversational agent use data from a synced source, not just repository files?

Yes. It can combine files already in the repository with synced source data to write a consolidated or updated document.

Where do I talk to the conversational agent?

In the Context tab, click the chat icon in the top right corner and type your request. You can resume past conversations from the same dropdown.

Can I stop the conversational agent mid-response?

Yes. The chat dropdown lets you stop an in-flight response, start a new chat, or go back to a previous one.

Make your company AI-native.

Make your company AI-native.

ISO 27001-certified

Enterprise-level security

GDPR-compliant

Hosted in the EU

Legal

© 2026 Qontext GmbH.

Legal

© 2026 Qontext GmbH.

ISO 27001-certified

Enterprise-level security

GDPR-compliant

Hosted in the EU

Legal

© 2026 Qontext GmbH.