Processors

A processor is one job: transcribe the audio, place b-roll, burn captions, adjust volume. A step is one processor with your settings filled in, and a template is an ordered list of steps.

Every processor here is available in the editor's Library tab, grouped by the same categories, and every page below has an In the editor section covering how to add the step, what to fill in, what to wire it to, and what you see after the run. The editor guide covers the surrounding app: adding steps, the Inspector, wiring inputs, and the YAML tab.

Each page also carries the YAML for the same step, the full parameter table, and notes for anyone writing templates as text.

Every step keeps its output so a later run can reuse it (see Reusing work between runs). Steps marked ✓ review can pause the run for your approval (see Review gates); in the editor that is the Review gate checkbox on the step.

When you add a new processor, add a row here and create docs/processors/<name>.md.

Transcript

Processor	What it does	Review
`generate_transcript`	Whisper transcription
`transcript_aiprep`	Prepare transcript text for TTS
`workspace_analyze`	AI summary of all project files
`assemble_transcript`	Merge per-clip transcripts into timeline transcript

B-Roll

Processor	What it does	Review
`broll_place`	Place existing clips in gaps	✓
`broll_veo_generate`	Generate clips with Veo 3	✓
`broll_imagen_generate`	Generate images with Imagen + motion	✓
`broll_websource`	Fill gaps with Wikimedia Commons media
`broll_screenshot_websource`	Animated web-source background clips

Transforms

Processor	What it does	Review
`reframe`	Scale landscape footage to portrait
`motion_center`	Pan to center subject (variance analysis)
`speaker_center`	Pan to center speaker (Claude Vision)
`chroma_key`	Green/blue screen removal
`scale_transform`	Scale and position patch
`edit`	Chain of edit operations (scale, color, crop, audio…)

Audio

Processor	What it does	Review
`elevenlabs_tts`	Text-to-speech via ElevenLabs
`audio`	Place/loop music or sound effects
`noise_reduce`	FFT audio denoising
`volume_adjust`	Gain adjust on a level
`sound_fx`	AI-placed social-media SFX on a dedicated audio level

Overlays

Processor	What it does	Review
`captions`	Per-word highlighted caption overlays
`title_card`	Rendered text overlays with fades
`drawing_list`	Animated bullet/numbered/checkmark list
`drawing_text`	Text/headline clip
`drawing_logos`	Company logos display
`drawing_diagram`	Animated flow diagram
`drawing_remotion`	Remotion composition (stats, chart, etc.)

Avatar

Processor	What it does	Review
`heygen_speaker`	Generate HeyGen avatar talking-head clips	✓

AI

Processor	What it does	Review
`ai_agent`	General Claude tool-use loop	✓
`write_script`	Generate narration script with Claude	✓
`video_director`	Produce a segment plan for downstream processors
`analyze_footage`	Analyze raw footage quality and content
`footage_assembler`	Auto-assemble timeline from analyzed footage	✓
`drawings_ai_placer`	AI agent places animated drawing clips	✓
`auto_broll_placer`	Unified AI agent: places existing broll, web media, screenshots, Veo clips, Imagen images, and drawing overlays in one session	✓
`media`	Generate or retrieve media from any external service via a spec

Utility

Processor	What it does	Review
`trim`	Cut a clip to in/out points
`cut`	Split a clip at timecodes
`stitch`	Concatenate files end-to-end
`deadspace_trimmer`	Remove silence using transcript	✓
`fill_level`	Background spanning full cut duration

Recording

Processor	What it does	Review
`av_recorder`	Pause pipeline to collect a live user recording