More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Parallax Source-to-Shorts

    1

    Turn your notes, interviews or product facts into three short-video scripts, with hooks, caption suggestions and filming notes.

    $7

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Parallax Source-to-Shorts

    Parallax Source-to-Shorts

    Example session with this skill installed

    Buyer request

    I am Alex, founder of FrameBrief, and I will speak in the videos. Use the material below to write three different English shorts for freelance brand designers in the UK who manage client feedback themselves. Aim for roughly 45-60 seconds each. Keep the tone calm and useful, with one light mention of FrameBrief where it fits. I want full spoken scripts plus practical filming and caption suggestions. Do not add a sales offer or a download. I own these notes and the product screenshots I can record.

    Source A: founder interview transcript

    Recorded for this fictional product's internal content session, 4 September 2026. Interviewer: Jo. Speaker: Alex.

    00:12 — Jo: What led you to build the review board?

    00:16 — Alex: On one packaging job I had feedback split between eleven emails and two image attachments. I changed the background colour, but I missed the requested change to the small subtitle. That was my mistake. I needed a way to see which requests were still open before I sent another version.

    00:49 — Jo: What do you look for when you review work now?

    00:54 — Alex: I separate a comment from a decision. “Maybe brighter” is something to discuss. “Use version B and keep the current subtitle” tells me what to do next. A board cannot make that decision for a client, but it can keep the decision next to the work.

    01:22 — Jo: Does that remove the need to talk to a client?

    01:26 — Alex: No. If two people want different things, I still need a conversation. I ask who is making the final call before I begin the next revision.

    Source B: FrameBrief product fact sheet

    Owner-maintained fact sheet dated 6 September 2026. Applies to the current beta described here.

    • B1. FrameBrief is a browser-based review board for image files and PDFs. It accepts uploads of up to 20 MB per file.
    • B2. Reviewers can place a comment on a particular image area or PDF page. A comment can be marked open or resolved.
    • B3. A project owner can label a version “in review”, “changes requested”, or “approved”. These labels are set by people; they are not automatic decisions.
    • B4. A review link lets invited reviewers view and comment without creating an account. The owner chooses which version the link opens.
    • B5. The beta does not include time tracking, automatic conflict resolution, or integrations with project-management tools.
    • B6. This fact sheet does not include pricing or comparative performance research.

    Source C: Alex's review handoff checklist

    Personal working method, written 5 September 2026. These are recommendations I use, not findings from a controlled study.

    • C1. Before sending work, write the decision you need from the reviewer. For example: “Please choose the cover image; we are not choosing typefaces in this round.”
    • C2. Name the person who will make the final call if comments conflict.
    • C3. Put the version name in the message and ask the reviewer to confirm that they are looking at it.
    • C4. Before starting a new revision, scan unresolved comments and confirm which requests belong in that revision.

    Available filming material

    A talking-head camera setup, a screen recording of the beta, and a made-up sample design for a fictional tea shop. The sample design is for demonstration and has never been a paying client's project. No client testimonials or measured outcome studies are available.

    • Read your context and instructions
    • Compiled the parallax source-to-shorts

    Fictional demonstration from an actual local trial. This is not a real customer result. Local links and machine-specific paths are displayed as text.

    Three scripts for Alex

    Audience: freelance brand designers in the UK who manage their own client feedback. Speaker: Alex, FrameBrief's founder. I used Sources A–C and the supplied filming options. British spelling and neutral UK-friendly phrasing are used; no prices, outcomes, offers, or downloads are added. The beta facts are owner-supplied and were not independently verified.

    ScriptViewer situationReason to keep watchingPayoff
    The tiny change you missedPreparing another design versionA visible fix can distract from a smaller unresolved requestScan open requests before sending the next version
    “Maybe brighter” is not a decisionReceiving vague feedbackActing on a comment can start a revision before a choice existsConvert the comment into a specific decision question
    Two reviewers, two directionsConflicting stakeholder feedbackA tidy board cannot tell you whose instruction winsName the final decision-maker and the version before revising

    Script 1 — The tiny change you missed

    I changed the background colour and still sent the wrong revision. The change I missed was a small subtitle, buried in feedback across eleven emails and two image attachments. That happened on one packaging job, and it was my mistake.
    
    What I needed was a way to see which requests were still open before I sent another version. That is why I built FrameBrief. In the beta, comments can sit on an image area or a PDF page, and you can mark them open or resolved.
    
    But the useful habit comes before the send button: scan the unresolved comments, then confirm which requests belong in this revision. The big visual change is easy to notice. The small instruction still needs a deliberate check.
    

    123 words; estimated 49–59 seconds, at 140–160 words/minute plus 3–6 seconds of pauses. Not a timed read.

    Film Alex talking to camera for the opening, then show the fictional tea-shop demonstration with a clearly labelled subtitle request. Do not present it as the real packaging job. Show a current beta open/resolved comment only if the supplied recording demonstrates it. End on the unresolved-comment scan. Caption moments: “One missed subtitle”; “Open or resolved”; “Check the requests in this revision.”

    Sources: A 00:16; B2, dated 6 September 2026 beta; C4. The first-person incident belongs to Alex. The tea-shop visuals are illustrative, not client evidence.

    Script 2 — “Maybe brighter” is not a decision

    “Maybe brighter.” Would you change the design, or ask what that means? If you start adjusting colours now, you could be revising before the client has made a choice.
    
    I separate a comment from a decision. “Maybe brighter” gives us something to discuss. “Use version B and keep the current subtitle” gives me something I can act on.
    
    Before sending work, write the decision you need. For example: “Please choose the cover image; we are not choosing typefaces in this round.” That gives the review a useful boundary.
    
    Then keep the agreed decision beside the work. A board can hold the answer, but it cannot choose for the client. When the comment is unclear, the next step is a clearer question, not another version.
    

    124 words; estimated 50–59 seconds, at 140–160 words/minute plus 3–6 seconds of pauses. Not a timed read.

    Film a talking-head take with two simple on-screen notes: “Maybe brighter” and “Use version B; keep subtitle.” These are example wording, not testimonials. Show two fictional cover options when Alex asks for a choice. Caption moments: “Comment ≠ decision”; “What choice do you need?”; “Ask before revising.”

    Sources: A 00:54; C1. The opening conditional consequence and final advice are interpretations of Alex's method, not measured performance claims. “Colours” follows the requested UK locale.

    Script 3 — Two reviewers, two directions

    One reviewer says, “Use version A.” Another says, “Use version B.” Putting both comments in one place does not tell you whose decision to follow.
    
    When people want different things, I still need a conversation. Before I begin the next revision, I ask who is making the final call.
    
    There is another check I make before that conversation: which version are we discussing? Put its name in the message and ask the reviewer to confirm they are looking at it.
    
    Those are two separate questions: are we discussing the same work, and who can settle the choice? Once those are clear, confirm the requests that belong in the next revision. A status label can record a decision. It does not make the decision for you.
    

    125 words; estimated 50–60 seconds, at 140–160 words/minute plus 3–6 seconds of pauses. Not a timed read.

    Film Alex beside the fictional tea-shop A/B designs. Label the opening as an example; do not invent a customer disagreement. Add a version name to a mock message and show a simple “Final decision: [agreed role]” teaching card without suggesting a new beta feature. If showing an actual beta status label, demonstrate a person setting it. Caption moments: “Same version?”; “Who makes the final call?”; “A label records a decision.”

    Sources: A 01:26; B3, B5; C2–C4. The illustrative opening is a hypothetical version of the documented conflict problem, not Alex claiming another real incident.

    Source and claim record

    IDClaim or adviceSource and locatorStatus and limitUse
    C01Eleven emails, two attachments, colour changed, subtitle missedA 00:16Alex's first-person account of one job1
    C02Needed to see open requests before sendingA 00:16Alex's own reason for building1
    C03Image-area/PDF-page comments can be open/resolvedB2Owner-supplied beta feature, not performance proof1
    C04Scan unresolved comments and confirm revision scopeC4Advice/personal method1, 3
    C05Separate vague comments from actionable decisionsA 00:54Alex's stated method; quoted examples exact2
    C06Ask for a particular decision, with a round boundaryC1Advice; exact cover/typeface example2
    C07Conflict still requires conversation and decision authorityA 01:26; C2Stated method, not automation3
    C08Name and confirm the versionC3Advice; no claim of a new product feature3
    C09People set beta status labelsB3; B5Current supplied beta fact; no automated conflict resolution3
    C10Vague comments may prompt premature revisionA 00:54 plus C1Conditional interpretation/advice, not a measured result2

    Pricing, performance comparisons, time saved, automatic decisions, and integrations were not added. FrameBrief is mentioned once by name. No unresolved source gap blocks these three spoken drafts. Product-screen filming still needs the actual beta capture; the writing does not assert that footage or captions have been rendered.

    Timing uses the actual script word counts and 140–160 spoken words per minute plus 3–6 seconds for pauses. This is a drafting estimate, not a timed read. Read Alex's takes aloud before recording, especially where a screen demonstration needs extra time. No video, voice track, or post was created.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Have useful material but no script to record? Turn your notes, interviews or product facts into three complete short-video scripts with different angles.

    What you get

    • Full spoken scripts with an opening that leads into the main point.
    • Filming notes, caption suggestions and estimated speaking times.
    • A source check showing where factual claims came from.

    Read a complete example: the source material, all three scripts and filming notes.

    Choose the audience, language, tone and target duration. Bring material you own or are allowed to use.

    Load this reusable skill in your Agensi-connected AI assistant. You receive scripts and filming guidance, not rendered videos or voiceover. AI service fees are separate.

    • Produces scripts and production guidance, not rendered videos, TTS, scheduled posts, or automatic publishing.
    • Does not independently verify supplied claims or guarantee views, retention, leads, or sales.
    • Speaking times are estimates that need a timed read; current platform specifications are not assumed.
    • Thin or conflicting material may support fewer useful scripts than requested. Different angles do not create new evidence.
    • Localization depends on the supplied audience and locale. The skill does not invent local laws, cultural preferences, currency rates, or market facts.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Permissions

    Read Files
    Write Files

    Use material you own or are authorized to adapt. The skill does not request account credentials, contact other people, access social accounts, or publish anything. Files are created only when requested and when the host assistant provides file tools. Content supplied to the host AI service remains subject to that service's terms and privacy settings. Text-only work stays in chat where supported; local files, calculations and previews use only the workspace and tools you authorize. No accounts are accessed or content published merely because a draft is ready.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    Paste an interview, your experience notes, a product fact sheet, or a combination. Add the intended speaker, audience, locale, and duration target if you know them. The skill can proceed with reasonable, visible assumptions when those details are not essential.

    Yes. Three scripts in international English is the default. Specify another count or language and any relevant country or locale. When a source is too thin, the output explains what can be supported instead of inventing additional facts.

    Each should serve a different viewer decision, problem, or useful takeaway. The concept map makes those differences reviewable before you record. Different wording around the same conclusion does not count as a distinct angle.

    It can use supported claims and accounts you supply. It will not invent your experience, customer praise, performance numbers, product features, or guarantees, and it identifies material facts that need confirmation.

    No. The one-time price is for the reusable skill. Your host AI service may charge separately. Video rendering, speech generation, editing, and publishing are outside this product.

    Say: Turn these notes into three different [45–60]-second videos for [audience]. The speaker is [person], the language is [language], and the tone is [tone]. Use only the facts below. Give me full spoken scripts, filming notes and caption suggestions. Then paste your notes, interview or product facts.

    New in Image & Video