Jobs & files
What a job is, how to read its status, where results and uploaded files live, and how to compare two runs.
Every time you run a tool, AgentLeverage creates a job to track it. A job holds the input you gave the tool, its current status, and — once it finishes — its output. Jobs are scoped to your organization, so anyone in it can see the same Job History.
Where jobs live
- Job History — every job you or your organization has run, newest first, with search and filters.
- A job's own page (
/app/jobs/<jobId>) — the full status, output, and actions for one run. Open it by clicking View on a completed or failed job in Job History.
Status
A job moves through one of these statuses:
| Status | Meaning |
|---|---|
| Pending | Queued, waiting to start. |
| Processing | Running now. |
| Completed | Finished successfully — output is ready. |
| Failed | Hit an error. You can run the tool again from the job's page. |
| Cancelled | Stopped before it finished. |
A job's own page auto-refreshes while it's Pending or Processing, so you can leave it open and watch it finish. You can only open the full output view for a Completed or Failed job — while a job is still Pending or Processing, Job History links straight back to the tool instead.
A job's page
Open any completed or failed job to see:
- Status — current state, when it was created, when it completed, and how long it took.
- Output — the tool's result, with a formatted view and a raw JSON view. Download
saves the raw output as a
.jsonfile, and Share copies a link to the job (anyone in your organization with access can open it). - Pin — pins the job to your dashboard for quick access later.
- Tags and Notes — free-form labels and a note you can attach to a job. These are saved in your browser only, so they don't follow you to another device or show up for teammates.
If a job Failed, its page explains what happened and links you back to the tool so you can try again — you'll need to re-enter your input, since it isn't carried over.
Who ran it
A job can be run by you in the app, or by an API token acting on your organization's behalf through MCP. Job History and a job's own page badge token-run jobs with the token's name so you can tell them apart from jobs you ran yourself. See Connecting your agent for how an agent's jobs relate to your own.
Finding a job
Job History supports:
- Search by tool name or job ID.
- Filters for status, tool, tag, and a date range.
- Bulk delete — select jobs with the checkboxes and delete them together.
- Export CSV — download the filtered list (ID, tool, status, timestamps, duration) as a spreadsheet. Available on paid plans; see Credits & billing for what each plan includes.
Comparing runs
Compare puts two completed jobs side by side so you can see how their outputs differ — useful for comparing two inputs, or the same input run twice. Pick one job for the left panel and one for the right; only completed jobs are listed. Like CSV export, this is available on paid plans.
Files
Files you upload — and files a tool needs as input, like audio for a transcription
tool — are tracked on the Files page for your organization
(/app/<your-org>/files).
- Upload by dragging a file in or clicking Upload Files. There's a 50 MB limit per file.
- Each file shows its type, size, and upload date, and you can attach tags to organize them — click a file's tags to edit them.
- Filter by file type (audio, image, video, document, other) or by tag, and search by filename.
- Download a file, or Delete it — deleting is permanent and removes it from storage.
Files don't expire on their own — they stay until someone deletes them.