> ## Documentation Index > Fetch the complete documentation index at: https://docs.bowerlabs.ai/llms.txt > Use this file to discover all available pages before exploring further. # Protocol runs > Every run of a protocol in your workspace, with the notes, photos, voice notes and recordings captured during it, newest first. Filter by protocol, source, collection, person, device and time. **Protocol runs** is the second tab on the [Protocols](/organisation/protocols) page. It lists every run of a protocol in your workspace, with the notes, photos and recordings captured during it, newest first. Open it to see what's running in the lab right now, or come back tomorrow to trace what happened on yesterday's run. This page used to be called the Capture Stream. The link `/capture-stream` still opens it, so bookmarks and shared filter links keep working. ## Opening Protocol runs Click **Protocols** in the left rail, then click **Protocol runs** beside the **Protocols** heading. You can also go straight to `/capture-stream`. To see the runs of one protocol, open that protocol's menu and choose **View captures**. Protocol runs opens already filtered to that protocol. See [Protocol menu and export](/organisation/protocol-menu-and-export). The page is workspace-scoped: you only see captures you have access to, gated by the same [permissions](/workspaces-and-account/permissions-overview) that apply everywhere else in Bower. ## Runs and captures The list under **Recent captures** holds two kinds of item. ### Run cards Captures taken while a protocol was running are grouped into one card per run. The card shows the run number (for example **Run 3**), the protocol it belongs to, when the run started, how long it took, and how far it got, such as **Step 4 of 12** or **All 12 steps done**. Click **Show captures** to expand the card and see what was captured during the run; click **Hide** to fold it again. A run with a single capture opens expanded. A card that isn't simply finished says so first: | Label | What it means | | - | - | | **Running now** | A live session is running this protocol right now. The card shows a running clock and updates while you watch. | | **Session ended** | The live session went away (a closed tab, a flat battery, a dropped connection) but the run is still open. Start the same protocol again within 30 minutes and it picks up at the step you were on. | | **Paused** | The run was paused. | | **Abandoned** | The run was stopped before it finished. | | **Never finished** | The run was left open and never closed. | A run that finished normally has no label, just its time range and steps. ### Single captures Anything captured outside a protocol run, such as a quick voice note from your phone or a photo of a label, appears as its own row: * **Icon and type** on the left: **NOTE**, **ATTACH** for an attachment, or **PROTOCOL**. * **Title and snippet** in the middle. For notes, the snippet is the start of the body. For attachments, it's the file's extracted text (OCR for images, transcript for audio, first paragraph for PDFs). * **Inline media** beneath the title. Images render as a thumbnail; click to open at full size. Videos show a poster frame; click to play. Audio files get an inline player, and so do notes that were transcribed from a recording, so the original audio is one click from the transcript. * **Time and origin** on the right. Hover the relative time ("3 hours ago") to see the exact local date and time. Below it, a source pill ("From audio", "AI written", "Live photo") tells you how the capture was made, then the collection it belongs to. ## Filters Six filters sit above the list. They all round-trip through the URL, so a link to *"Mish's captures during PCR setup, last 7 days"* is shareable. | Filter | What it does | | - | - | | **Source** | Multi-select. How the capture was made, such as **Voice**, **From audio**, **Image**, **Live photo**, **Live recording** or **AI written**. | | **Collection** | Multi-select. Only captures inside the chosen collections. | | **Protocol** | Multi-select. Only captures taken while running the chosen protocols. | | **Created by** | Multi-select. Only captures made by the chosen workspace members. | | **Device** | Multi-select. Where the capture came from: **Mobile**, **Desktop**, **Live agent**, **Bower chat**, **Connector** or **API**. | | **Time** | Today, last 7 days, last 30 days, or all time. | **Device starts on Mobile and Live agent.** Protocol runs is about work at the bench, so files you drag in on your desktop and imports through a connector or the API are hidden until you add them to the Device filter. If nothing shows yet, the empty state offers **Show all sources** to include everything. Click **Clear filters** to drop them all and go back to the default. ## Opening a capture and coming back When you open something from Protocol runs, the top of its page shows a trail back to the list, such as **Protocol runs / PCR setup**, with the collection it's in beside it. Click the back arrow to return to Protocol runs with your filters still applied. Closing or deleting the item also takes you back there. ## What's not in the list * **Nested attachments** under another note (for example, an image you dropped into a note's editor). Those belong to the parent note. * **Trashed items.** Restoring a trashed capture brings it back automatically. * **System-internal items**, such as the retired per-step protocol annotations. ## Related docs * [Protocols](/organisation/protocols): the library on the tab next door * [Live agent mode](/ai-and-search/live-agent-mode): running a protocol hands-free, which is where run cards come from * [Mobile capture](/capture/mobile-capture): voice and photo capture from your phone * [Batch import](/capture/batch-import): bringing existing files into Bower * [Collections](/organisation/workspace-items#collections): how captures get organized This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.