Chat
Chat is the fastest way to get something done in JobKred. You describe what you want in plain language, and Ellie — your AI assistant — figures out which specialist to use, does the work, and returns a result you can read, edit, and export.
Each conversation is a chat. A chat keeps its full history, the context you've attached, and every result Ellie has produced, so you can always come back to it.
How a conversation works
- You send a message. It appears in the chat immediately, and the input shows "Agent is processing your request…".
- Ellie works in the background. The work runs on our servers, not in your browser, so it keeps going even if your connection drops or you switch tabs.
- The answer appears in the chat when it's ready — usually within a minute, longer for larger analyses.
Because processing isn't tied to your browser tab, you can navigate away — even close the app — and the answer will still be waiting in the chat when you return.
Stopping a response
If you start a response and realize you don't need it, leaving the chat prompts you to cancel it. Cancelling stops the work so nothing keeps running in the background.
How Ellie decides what to do
You never have to pick which tool or specialist to use — Ellie decides for you. For every message you send, Ellie does one of three things:
| Outcome | What you see |
|---|---|
| Answers or delegates | Ellie either replies directly or hands your request to the single best specialist agent for the job, then returns its result. |
| Asks for clarification | If your request is vague, incomplete, or needs a document you haven't attached, Ellie asks a follow-up question instead of guessing. |
| Declines | Ellie won't act on requests that fall outside its purpose — off-topic, unsafe, or attempts to make it ignore its instructions. |
Ellie deliberately tackles one task per message. If you have a multi-step goal, send it as a sequence of messages — or build a workflow to chain the steps together automatically.
To see the full range of specialists Ellie can call on, browse the Agent Catalog.
What's in this section
- Start a conversation — starter suggestions, what to ask, and follow-ups.
- Add context — ground answers in your frameworks, files, and web links.
- Read, export & sync results — preview, save, download, and push outputs to connected systems.
- Manage chats & projects — history, projects, tags, sharing, and workflow-linked chats.