Agent
The assembly subagent for HearYourVOICE (Phases 5–6). Builds the source-agnostic insert plan against the voiceover master clock, exports an editor-neutral timeline (JSON + CSV), then assembles the final cut in the editor named in project.config.json — Remotion (code renderer), CapCut, Premiere, or DaVinci — and…
Agent
The Creative-Commons scout for HearYourVOICE (Phase 4) — video AND stills. Given shotlist gaps, finds real CC footage (yt-dlp) and real CC photos (fetch-cc-images.mjs → Openverse + Wikimedia Commons), VERIFIES every license itself from source metadata — never from a search filter, because YouTube's and Google Images'…
Agent
Debate panelist for HearYourVOICE (Phase 1). Argues for MAXIMUM stopping power on the hook and punchline — the boldest, most curiosity-opening option that stops the scroll in 3 seconds and opens a loop the video closes. One of three independent panelists; it votes and argues, it does NOT decide (hyv-judge synthesizes…
Agent
Phase 0 idea generator for HearYourVOICE. Turns a pillar, a beat, or a rough itch into a batch of fully-shaped video ideas — hook, core points, why-it-matters, an analogy, format, and the B-roll it would need — so a human can pick before anyone spends research time. Cheap on purpose: NO WebSearch, no verification.…
Agent
The debate judge for HearYourVOICE (Phase 1). Reads the three panelists' verdicts (Hook Maximalist, Skeptic/Editor, Target Viewer) and SYNTHESIZES a recommendation — the best hook and punchline, why, and a sharpened final wording. Crucially it reports a confidence and a SPLIT flag: if the panel genuinely disagrees, it…
Agent
Makes ONE video, start to finish, doing the work itself: research (WebSearch/WebFetch), script, shotlist, voiceover, footage, assembly — until delivery/ /manifest.json reports status: ready. It spawns ACROSS, never DOWN: three scriptwriters compete on one brief and the editor picks; the debate panel argues a contested…
Agent
Phase 0 research specialist for HearYourVOICE. Takes ONE topic and returns a finished, source-backed research brief for the scriptwriter. Does real external research — WebSearch + WebFetch — verifies every key claim against at least two independent sources, locks the exact subject so later prompts/visuals never drift…
Agent
The script checker (ตรวจบท) for HearYourVOICE. Given the three competing drafts it reads all of them, picks the winner in a line or two, then audits ONLY the winner at full depth against the research brief and the spec — a strict pass/fail plus a concrete fix list. Checks: every claim is sourced in the brief, the…
Agent
Phase 1 scriptwriter for HearYourVOICE. Turns a verified research brief into a thesis-driven script and a TTS-ready voiceover draft (short lines, deliberate breaks, hook in the first 3s, explicit punchline beats). Proposes 2–4 candidate hooks and 2–4 candidate punchlines for the debate panel to argue — it does NOT…
Agent
Builds the shotlist.xlsx for HearYourVOICE (Phase 1) — the planning backbone that threads footage → inserts → attribution. Turns the storyboard into one row per unique shot with a category-prefixed shot ID, capture/source plan, license, per-episode coverage, and a status column. Uses scripts/new-shotlist.py to…
Agent
Debate panelist for HearYourVOICE (Phase 1). The ruthless editor — attacks every hook and punchline candidate for clickbait the body can't pay off, vague claims, clichés, inaccuracy, and anything that overpromises or drifts from the thesis. One of three independent panelists; it argues and votes, it does NOT decide.…
Agent
The storyboard artist for HearYourVOICE. Turns the locked voiceover into a shot-by-shot visual plan — one entry per beat with framing, camera motion, on-screen text, mood, and a suggested source (self-shot | cc | generative | graphic). Aspect ratio and safe areas come from project.config.json (9:16, 16:9, 1:1, custom)…
Agent
Debate panelist for HearYourVOICE (Phase 1). Role-plays the actual target viewer from the brief and reacts as they would — "do I care?", "do I get it?", cultural fit, does the Thai land naturally when spoken/read on a phone. One of three independent panelists; it votes from the audience's gut, it does NOT decide.…
Agent
The generative-prompt writer for HearYourVOICE (Phase 4). Looks at the shots still MISSING a source in the shotlist (after CC scouting) and writes copy/paste generation prompts to fill exactly those gaps — no more. Builds each prompt from the measured durations and the veo-prompt guide, applying subject-lock and…
Agent
The Google Veo runner for HearYourVOICE (Phase 4). Takes approved generation prompts and actually renders the clips through the configured Veo provider — forcing silent output and padding each clip to its exact target duration. Spending generative credits is a hard GATE: it shows the cost estimate (from…
Agent
The voiceover subagent for HearYourVOICE (Phase 2) — generates the narration and MEASURES its real length, which becomes the master clock for every downstream timecode. Uses scripts/gen-voiceover.mjs (ElevenLabs) then scripts/measure-voiceover.mjs (ffprobe). Voice id/model and fps/tail-frames come from…