Flow Kit Docs
34 skills, 10 core concepts, and battle-tested workflow recipes. Search or browse — every card links back to what the underlying API call does.
Core Concepts
Ten ideas that explain how Flow Kit stays consistent, recovers from failure, and produces YouTube-ready output from a single project definition.
Reference Image System
One image per entity, reused in every scene it appears
Every visual element that should stay the same across scenes — a character, a location, a prop — is declared as an entity. Flow Kit generates exactly one reference image per entity and passes that image as imageInputs to every scene that lists the entity in character_names. Identity does not drift because the same reference is seen by the model every time.
character_names: ["Pippip", "Fish Stall", "Open Market"] → scene image is composed using all three reference images.
Entity Types
character · location · creature · visual_asset
Entity type drives the aspect ratio of the reference image. Characters, creatures, and visual_assets get portrait refs (full body, front-facing, centered). Locations get landscape refs (establishing shot, level horizon). Faction and generic_troop are available for multi-character groupings.
Chain Types
ROOT → CONTINUATION → INSERT
Scenes form a tree. ROOT scenes start a sequence. CONTINUATION scenes link to a parent and reuse its end frame as their start frame for smooth transitions. INSERT scenes branch off a parent for cutaways, close-ups, or multi-angle shots without breaking the main chain.
Two Prompts Per Scene
prompt (still) vs video_prompt (motion)
prompt describes the frame-0 still image. video_prompt describes the 8-second motion with sub-clip timing (0-3s / 3-6s / 6-8s), camera directions, dialogue, and sound effects. Keep scene prompts as action — never repeat entity appearance (the reference image carries that).
prompt: "Luna steps out of rocket onto candy planet. Wide shot, sunrise." video_prompt: "0-3s: wide crane down, Luna gasps. 3-6s: low angle tracking, shallow DOF. 6-8s: close-up, golden hour backlight."
Media ID = UUID
Never a CAMS... string
Every media_id is a UUID (xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx). The worker refuses to start scene generation if any referenced entity has a CAMS... mediaGenerationId. /fk-fix-uuids extracts the UUID from fifeUrl and backfills the right value automatically.
Dual Orientation
VERTICAL (9:16) + HORIZONTAL (16:9) in one project
A scene table carries vertical_* and horizontal_* columns so a single project produces both a YouTube Short and a long-form video without duplicating entities, prompts, or narrator text. Upload auto-detects Shorts (under 61s + 9:16) vs long-form.
Material System
realistic · 3d_pixar · anime · stop_motion · minecraft · oil_painting
Every project picks one material. The material controls both entity image_prompt style and the scene_prefix applied to every scene prompt, so visual style stays coherent across references and scenes without per-prompt repetition. Custom materials can be registered.
Voice Description
Max ~30 words per character
Characters carry a short voice_description that the worker auto-appends to every video_prompt where they appear. Keep it concise — tone, pitch, cadence, accent. Combined with the project voice template, it keeps speaking performance consistent across scenes.
i2v / i2v_fl / r2v
Three ways to get from pixels to motion
i2v animates a single start frame. i2v_fl chains start + end frames for smooth CONTINUATION transitions controlled by transition_prompt. r2v builds a clip from entity references only (no scene image) — ideal for intros, dream sequences, or montages.
TIER_ONE vs TIER_TWO
Upscale + Veo 3.1 relaxed live on TIER_TWO
Check tier with GET /api/flow/credits. TIER_ONE has access to standard Veo 3.1 image + video generation. TIER_TWO unlocks 4K upscale (UPSCALE_VIDEO request) and relaxed safety filters. The worker auto-downgrades models when it sees MODEL_ACCESS_DENIED.
Suggested Workflows
Pre-built skill sets for common video production scenarios. Pick a recipe and run the commands in order.
Quick Video
BeginnerThe fastest path from idea to a finished clip. Create a project, generate refs + images + videos, and concat into a single output.
Self-Healing Pipeline
BeginnerRun the full pipeline end-to-end and let /fk-doctor triage any Flow, extension, or worker error automatically before retrying.
YouTube Ready
IntermediateComplete pipeline from script to published YouTube video with narrator TTS, SEO metadata, branded thumbnails, and rule-validated upload.
Chain Continuity
AdvancedCinematic chained sequence using i2v_fl start+end frame transitions, mid-chain inserts for cutaways, and a vision-reviewed quality gate.
Documentary
AdvancedResearch-backed documentary with fact-checked scripting, narrator voiceover, data overlays, and music — safe against real-people safety filters.
Shorts Factory
BeginnerRapid vertical Shorts production. Vertical-only orientation, minimum steps, optimized for batch publishing with channel rules.
All Skills
Every /fk-* skill available in Flow Kit, grouped by what it actually does for you.
Core Pipeline(5)
/fk-create-projectCreate Project
Interactive wizard that sets up a project with entities (characters, locations, assets), dual-orientation videos, and scenes with chain_type ROOT/CONTINUATION.
/fk-gen-refsGenerate References
Generate one reference image per entity (portrait for characters/assets, landscape for locations). Verifies all responses return UUID media_id before proceeding.
/fk-gen-imagesGenerate Scene Images
Generate frame-0 images for every scene with all referenced entities applied via imageInputs. Blocks if any required ref is missing a UUID media_id.
/fk-gen-videosGenerate Videos
Animate scene images into 8s video clips via Veo 3.1 (i2v). Polls until complete. Auto-appends voice_description and 'no background music' rules to prompts.
/fk-concatConcat Final Video
Download every scene video, normalize with ffmpeg, and concatenate into a final cut. Preserves sound effects, no background music by default.
Advanced Video(6)
/fk-gen-chain-videosChain Videos (i2v_fl)
Generate videos with start+end frame chaining for smooth transitions between CONTINUATION scenes. Uses transition_prompt to control visual smoothness.
/fk-insert-sceneInsert Scene
Insert cutaways, close-ups, or multi-angle shots into an existing chain as INSERT children. Maintains ROOT/CONTINUATION integrity.
/fk-creative-mixCreative Mix
Analyze your story and suggest mixed techniques: chain transitions, inserts, r2v intros, parallel multi-scene generation for cinematic impact.
/fk-review-videoReview Video
Claude Vision quality pass before upscale: flags artifacts, continuity breaks, motion glitches, and off-brief compositions for regeneration.
/fk-review-boardReview Board
Visual scene review board for bulk feedback. Flag scenes for REGENERATE_VIDEO or REGENERATE_IMAGE with targeted notes, batched in one pass.
/fk-add-materialMaterial Style
Register or switch visual material: realistic, 3d_pixar, anime, stop_motion, minecraft, oil_painting, or custom. Controls both image_prompt and scene_prefix.
TTS & Narration(6)
/fk-gen-tts-templateVoice Template
Create a reusable OmniVoice anchor WAV for consistent narration. Zero-shot voice cloning across 600+ languages. Re-used per scene at generation time.
/fk-gen-narratorGenerate Narrator
Auto-write per-scene narrator_text, then run TTS using the project voice template. Skips interview scenes and preserves original audio when flagged.
/fk-gen-text-overlaysText Overlays
Extract dates, locations, stats and callouts from narrator text and schedule timed text overlays for the final cut.
/fk-concat-fit-narratorConcat Fit Narrator
Trim every scene video to match its narrator TTS duration with xfade cross-dissolve on chains, burn text overlays, and concat at 48kHz audio.
/fk-gen-musicGenerate Music
Generate a Suno chirp-v4 soundtrack matched to video mood. Uses sunoapi.org. Only applied when project.allow_music is enabled.
/fk-import-voiceImport Voice
Register an existing WAV as a project voice template without re-recording. Handy for reusing voice anchors from prior projects.
YouTube(4)
/fk-youtube-seoYouTube SEO
Generate niche-aware title, description, hashtags, and tag list. Validates total tag-char budget (<=500 with quote overhead) to avoid invalidTags.
/fk-brand-logoBrand Logo
Overlay channel logo intro/outro (220px or 4K badge from channel directory) on final video and thumbnails, covering the Veo watermark.
/fk-youtube-uploadYouTube Upload
Upload via YouTube Data API v3 with OAuth2. Auto-detects Shorts vs long-form, enforces per-channel max-per-day, min-gap, and avoid-hour rules.
/fk-thumbnailThumbnails
Generate 4 AI thumbnail variants optimized for CTR with branding and text overlays. Pulls channel design rules from channel_rules.json.
Orchestration(5)
/fk-pipelineFull Pipeline
Auto-detect current project state and run the remaining stages in order. Uses POST /api/requests/batch so the worker handles throttling (5 concurrent, 10s cooldown).
/fk-monitorPipeline Monitor
Poll project pipeline state, detect transitions, send desktop notifications, and optionally auto-download completed upscales.
/fk-statusStatus Dashboard
Print a full project dashboard: scene progress, entity refs, pending requests, error history, and the recommended next skill to invoke.
/fk-switch-projectSwitch Project
Switch the active project across the agent so subsequent skills target the right slug automatically. Reads/writes the active-project API.
/fk-dashboardLive Dashboard
Render a live Flow Kit statusline in Claude Code so progress, failures, and retries are visible inline while you work.
Diagnostics(3)
/fk-doctorDoctor
Auto-diagnose pipeline errors across Flow backend, Chrome extension, worker, and YouTube upload. Prescribes targeted fixes — invoke before guessing.
/fk-fix-uuidsFix UUIDs
Find media_id values stuck in CAMS... form, extract the real UUID from fifeUrl, and backfill scenes + entities so downstream gens succeed.
/fk-refresh-urlsRefresh URLs
Recover expired GCS signed URLs for scenes and entity refs (roughly 1h TTL). Handles the 'Requested entity was not found' recovery path.
Utilities(2)
/fk-change-modelChange Model
Inspect and switch video/image/upscale models per project via models.json. Downgrades automatically if tier does not have access.
/fk-upload-imageUpload Image
Upload a local PNG/JPG to Google Flow and receive a UUID media_id. Use for bespoke scene images, manual refs, or brand assets.
Reference(3)
/fk-camera-guideCamera Guide
Reference for cinematic video prompts: camera angles, movements, lighting, depth of field, and shot-timing patterns for 8s clips.
/fk-thumbnail-guideThumbnail Guide
Design rules for CTR-optimized YouTube thumbnails: composition, color psychology, hook phrases, and required branding elements.
/fk-researchResearch
Fact-check dates, names, and events before writing documentary content. Pulls citations and flags real-people bypass rules for safety filters.