> ## Documentation Index
> Fetch the complete documentation index at: https://docs.blindsight.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Projects

> Group datasets, scans, schedules, and integrations into the same logical workspace. Pin them, share them, and bulk‑manage from one screen.

A **Project** is the top‑level container in Blindsight. Every dataset,
scan, report template, integration setting, and schedule lives inside
exactly one project. Use projects to mirror your team structure or
your initiative structure: one project per model you're shipping, one
per dataset family, one per regulated domain, whatever fits.

## What a project gives you

* **Scoped membership** beyond the workspace‑wide roles, a user can be
  a workspace `Member` and a project `Admin`.
* **Per‑project integrations** so a clinical project can point at a
  different S3 bucket than a public dataset project.
* **Per‑project report templates** so compliance reports come out
  consistently branded for each domain.
* **Per‑project schedules** so you can re‑scan an entire project on a
  cadence with a single cron entry.

## Project list (`/projects`)

The project list is the entry point for the section. Each card shows
the project name, a brief description, severity rollup, and the last
activity date.

<Steps>
  <Step title="Create a project">
    Click **New project**, pick a name and a description, optionally
    set an icon and color. The creator becomes the project owner.
  </Step>

  <Step title="Pin frequently used projects">
    Click the pin icon next to any project. Pinned projects float to
    the top of the sidebar across the whole app.
  </Step>

  <Step title="Invite members">
    From the project detail page, **Members → Invite**. Pick a role
    (see [Team & Access](/data-integrity/team)) and Blindsight sends an
    invite email.
  </Step>

  <Step title="Archive when done">
    Archived projects disappear from the sidebar but stay queryable
    in compliance reports and the audit trail. Archive instead of
    deleting whenever auditability matters.
  </Step>
</Steps>

## Project detail page (`/projects/:id`)

When you open a project, you land on a dashboard tailored to it.

| Section              | What it shows                                                       |
| -------------------- | ------------------------------------------------------------------- |
| **KPI strip**        | Dataset count, scan count, critical findings, last activity.        |
| **Dataset list**     | Every dataset in the project with severity chips and quick actions. |
| **Recent activity**  | Timeline of scans, uploads, healings, and member changes.           |
| **Report templates** | Project‑scoped templates that override workspace defaults.          |
| **Schedules**        | Cron entries that scan datasets in this project.                    |
| **Members**          | Users with explicit access, plus their project role.                |

## Common workflows

<AccordionGroup>
  <Accordion title="Set up a project for a regulated workload">
    1. Create the project with a clear name (`clinical-radiology-v2`).
    2. Under **Members**, add only the people with PHI clearance and
       set workspace `Viewer`s up to `Reviewer` inside the project.
    3. Under **Integrations**, point the project at the right S3
       bucket and configure a project‑local report template branded
       for the auditor.
    4. Add a schedule that runs the relevant engines weekly and emails
       the team on completion.
  </Accordion>

  <Accordion title="Hand a project off to a different team">
    1. Promote the new lead to project `Admin`.
    2. Make sure they own at least one workspace API key with the
       right scopes (see [Team & Access](/data-integrity/team)).
    3. Demote the previous lead to `Reviewer` if you want them to
       retain read access, or remove them entirely.
    4. Archive any datasets or schedules that the new team won't keep.
  </Accordion>

  <Accordion title="Bulk‑clean a project">
    Use **Bulk operations** on the project detail page to delete every
    dataset and scan in one click. This is destructive and audited;
    confirmation is required, and the action surfaces under
    `dataset_action_logs` in the [audit trail](/data-integrity/audit-trail).
  </Accordion>
</AccordionGroup>

<Warning>
  Project deletion is permanent and cascades to every dataset, scan,
  and result inside the project. Archive instead if you might need
  the data for compliance later.
</Warning>