> ## Documentation Index
> Fetch the complete documentation index at: https://docs.get-hive.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Workflow assistant
> Draft a Hive workflow from a plain-language description and change it through chat, with every proposed change waiting for your review.
The workflow assistant drafts workflows from a description and proposes changes to the one you
have open. It never changes a workflow on its own: a new draft does not run until you deploy it,
and a change to an existing workflow waits for you to apply it.
## Draft a workflow from a description
On the **+ Describe** tab, write what should happen — the trigger, the steps, the apps involved —
or pick a starter brief, then select **Generate workflow**. Hive creates a new private draft and
opens it in the **Builder**. The draft records where it came from: a built-in template or the
model, and which model.
Name the trigger and the apps. "Every weekday at 9, find overdue Xero invoices and draft a polite
reminder in Outlook for each" gives the assistant far more to work with than "chase invoices".
## Open the chat
In the builder, select the chat button beside the canvas. On a new, empty workflow it reads
**Create a new workflow**; on an existing one, **Change this workflow**. The chat shows the model the
assistant uses, or **No AI model configured**. Only members who can edit the workflow can ask it to
change it.
## Change a workflow
Describe what should be different. When the assistant proposes a change, the reply shows a
**Proposed change**: a summary (up to 500 characters) and a preview of the steps it adds, removes
or edits.
* **Apply proposed change** saves your pending edits, then applies the proposal. If the workflow
changed after the proposal was made, Hive reports the conflict instead of overwriting the newer
version.
* **Dismiss** leaves the workflow as it is.
A reply without a proposal answers your question and changes nothing.
## What the assistant checks
Every draft and proposal is checked the same way a deploy is: the steps must be valid, allowed for
your role and runnable in your workspace. When a draft fails that check the assistant gets one more
attempt; if that fails too, nothing is created or changed and the chat says why.
When the assistant drafts a loop that could exceed the
[500-step run budget](/workflows/runs-and-errors#the-500-step-run-budget), it lowers the loop's
**Max iterations** to fit and says so, for example "Capped the loop at 166 items so a run stays
within Hive's step limit." A workflow you build by hand is refused instead.
The assistant never adds a **Saved agent** step; add that yourself.
## Conversations
Each workflow keeps its **Original conversation**. **New conversation** starts another, created when
you send its first message. **Chat history** lists earlier ones (up to 50 per workflow).
## Attach files
Select **Add files** to send files with a message:
| Limit | Value |
| - | - |
| Files per message | 3 |
| Size per file | 1 MB |
| Types | TXT, Markdown, CSV, JSON, PDF, DOCX |
| Text read per file | 20,000 characters; longer text is cut and the file is marked as cut |
| Message length | 10,000 characters |
Reading attachments needs a configured AI model.
## Preferences
**Chat preferences** opens **Workflow Chat Preferences**: list your **Preferred Tools** per category
and add **Custom Instructions** (up to 2,000 characters) about behaviour, style and tone. Only your
own messages use your preferences.
## Without an AI model
If no AI model is configured for your workspace, the assistant can only start a draft from a
built-in [template](/workflows/templates) whose starter brief you send unchanged (case and spacing
do not matter). Its reply reads "I prepared the … template as a proposed change." Anything else is
refused.
## Related
Other ways to start, and editing the canvas.
The built-in starting points.
How your workspace chooses AI models.
Check the draft before it goes live.