The capture dock — adding a finding

What this screen is for. The bar at the bottom of the capture screen is how a finding is born: say what you see, snap the photo, tap one button — the AI names the component, picks the section, rates it, writes it up, and files it. How to get there. It's pinned to the bottom of the capture screen. If it's tucked away, tap the "Add a finding" handle to bring it out — tucking it never throws away a note or held photos. What every button does Talk — the big one. Tap it and say what you see, the way you'd tell the client: "tub spout is loose and the caulking's failed." Tap Stop when done. While the mic is live a red-dot notice stays on screen: "Recording — tell anyone in the home" — a reminder to let people present know, because some states require everyone's OK. You can type into the note box while it listens — typing is never erased by speech. Photo — on a phone, one tap opens the camera, period. On a computer it asks where the photo is coming from: Take a photo, 360° photo, or From your photos. Held photos line up in a strip — rotate, Mark up (arrows and circles burned in), or remove them before saving. Video — an in-app recorder that hears you: whatever you narrate lands in the note, and the clip plays on the finding and in the report. While it films, the same red-dot notice shows — "Recording — tell anyone in the home" — with one more line saying the clip keeps the room's sound. Tap Mute sound before you record and the video is filmed with no audio at all (your narration still becomes the finding); the app remembers your choice for next time. Light — the flashlight, first in the row so a hand in a dark crawlspace finds it. It stays on when you take a photo. Help me — ask about what you're looking at: snap a photo and say "what year is this water heater?" or "is this pipe copper or PEX?". The sheet that opens is headed Ask HomeRecon Genie — same helper, tap-to-ask by voice, text, or photo. When it answers, one tap adds the answer to the report as a finding. The dock mic pauses while it's open and resumes after. Add manually — a plain form (section, component, location, rating, what you saw, what you recommend) for when you'd rather write it yourself. A manual finding is yours as written — there's no AI review step on it. Comment — your saved phrases. With an empty note, tapping a comment's circle starts a finding from it (you'll see "From your comment: …"); with a note in progress, it attaches to that finding. Tap a comment's name to read it first; this section's comments come up first, most common finding first. There's also Save this note as a comment to bank a good sentence for next time. Phone photos (phone only, second row) — a photo already on this phone, or a 360° photo. That's where 360° lives on a phone; on a computer it's inside the Photo menu. For a 360° you pick a pano file you've shot; the sphere renders on the finding and in the report. Analyze w/ AI — the green button that files it. Its label tells you what it's about to read ("Analyze 2 photos & text w/ AI"), and it reads "Writing it up…" while the AI works. A placeholder row shows the finding being written, then the real tile appears — in the section the AI chose, wearing a "Review AI finding" pill until you review it. Save without AI — right beside Clear, whenever there's a note or a photo. Same finding, filed the same way, but without running the AI: it takes your words (a saved comment, say) and any photos exactly as they are and files a plain finding with a Manual badge. For when you just want to tap a comment, add a picture, and move on — no waiting, no AI write-up. A video with nothing to read has only the one green save (there's no AI to skip). Common questions "How do I add a finding, start to finish?" Talk → Photo → Analyze w/ AI. Three taps. The AI writes it up and files it; you review it before it ever reaches a client. "Why does it say 'Recording — tell anyone in the home'?" Courtesy and law: some states require everyone's consent to record. It shows whenever the mic is live — on Talk and on Video — and goes away when you tap Stop. "Can I take a video without recording the room's sound?" Yes. On the video recorder, tap Mute sound before you start. The clip is filmed with no audio at all, which is the safe choice in states that require everyone's consent to record. Your spoken notes still become the finding, and the app keeps your choice for the next video. "Where's the 360° button?" Phone: Phone photos → 360° photo. Computer: Photo → 360° photo. "What if the AI is unreachable mid-house?" Your finding still files, instantly, from your own words — you'll see "Saved your note — the AI couldn't be reached; write it up yourself," and the finding carries a Draft this finding with AI button for when it's back. Nothing you say ever disappears. "I fired off five captures in a row — do they collide?" No. They file one at a time, in the order you tapped — "Filing:" on the current one, "Queued:" on the rest — and a stuck one never blocks the line. "Can I start a finding from one of my saved comments?" Yes — Comment with an empty note. The comment's words drop into the note and file like anything you'd have said. "Do I have to talk? I'd rather type." Type straight into the note box, or use Add manually for the full form. Good to know The finding lands where it belongs, not where you're standing. The AI picks the section from what you said — a water-heater note files under Plumbing even if you tapped from the Garage. The Findings chip up top shows you where everything went. On a phone the capture row is exactly five squares — Light · Photo · Talk · Video · Help me — with Add manually, Comment, and Phone photos on the row below. On a computer it's one row with Add manually and Comment at the edges. On a tablet (one wide row, but Photo opens the camera directly like a phone), Phone photos sits right beside Photo, so a photo already on the device is one tap away there too. The layout is pinned by tests on purpose — muscle memory matters mid-house. Video's steadier-camera option: if Video is set to use the phone's own camera app, a note reminds you that recording in-app is what transcribes your narration. Guided 360° shooting isn't built yet — today you pick a 360° file you've already captured with your pano camera or app. Related The capture screen — the shell around this bar A finding — what the AI wrote, and editing it Comments — the phrases behind the Comment button The checklist — ticking lines and aiming the dock