Beginners don't need more tools: they need a loop. A workflow is just a repeatable sequence: plan the video, generate the shots, assemble the edit, finish with sound. Run the same loop ten times and you're intermediate. Here's a version sized for a weekend.
Phase 1: Plan (30 minutes)
Write one paragraph describing the video: who it's for, the hook, the middle, and the ending. Then break it into 3–5 shots, one sentence each. This becomes your shot list: the document every later decision refers back to.
- Concept paragraph: audience + hook + payoff
- Shot list: 3–5 shots, one sentence each
- Note the visual style in three words (e.g. “soft, warm, minimal”)
Phase 2: Generate (60–90 minutes)
Work through the shot list in order. For each shot, write the prompt using the four-part frame: subject, action, camera, light: generate, judge against the plan, and iterate up to three times before moving on. Save every prompt with its clip.
Resist the urge to perfect shot one while shots four and five don't exist. A complete rough set beats one perfect clip every time.
Phase 3: Assemble (60 minutes)
Import everything and build a rough assembly first: shots in order, trimmed to their best seconds, no effects. Watch it twice. Cut anything that doesn't serve the hook-to-payoff spine: usually 20–30% of what you generated.
Only then add transitions (sparingly), adjust pacing, and check the video still makes sense with the sound off.
Phase 4: Finish (30–45 minutes)
Add voice or captions first: they carry meaning: then music and effects underneath at low levels. Export for your target platform, watch the export on a phone, and fix anything that only appears at small size (tiny text, muddy audio).
- Captions readable on a phone screen
- Voice clear, music at least 10dB under
- Hook visible in the first 2 seconds
- File named, prompt log saved with the project