YouTube Automator

Connections

Key Metrics

—

Loading…

Base Summary

Jobs

ID	Type	Ch	Status	Progress	Queued	Updated	Result

Downloads

Click Refresh to load downloads.

Each form creates a queue job. Pick the API key (tag) and model from your API Rotator; the worker dispatches to the right connector. Models are fetched live from your Alibaba key.

Video context (auto-fills defaults + save folders)

Channel

Project / Series

Video

Pick a channel to auto-fill tool defaults; pick a video to route outputs into its project folder.

Social Calendar

Multi-platform visual planner & scheduling calendar. Authorize mock accounts, schedule posts with text or local assets, and trigger manual force-publishing (Buffer alternative).

Resource Telemetry

Monitor active API token consumption across cloud endpoints and local GPU execution amortization. Manage budget alerts and simulate expenditures (Hootsuite alternative).

Teleprompter

Scrolling script reader + voiceover recorder — narrate while you read; the take is saved as .wav into your chosen output folder (queued, shown on the Queue tab). The same teleprompter from Make Video, standalone.

Channel (optional)

Script file (.md / .txt — loaded into the teleprompter; optional)

Output folder (where the recording .wav is saved)

00:00

Webcam recorder

Record an on-camera A-roll take in the browser, preview it, then save the video into your output folder (e.g. 02_Footage/A_Cam). The same webcam recorder from Make Video.

Output folder

Audiobook captions

Turn a voiceover (audio) or a transcript/SRT into word-grouped Remotion captions and open them in the AI editor. The same audiobook-captions builder from Make Video.

Channel (optional)

Audio file (or paste a transcript/SRT below)

Transcript or SRT (used when no audio path)

Words per cue

Size

Output folder (blank = default animations folder)

Word-by-word captions

Word-by-word synced captions exported for DaVinci Resolve — writes a stylable .srt (File → Import → Subtitle) + a best-effort caption .fcpxml. Set words/cue = 1 for true word-by-word. Uses an audio file (transcribes word timings) or a pasted SRT.

Channel (caption template)

Audio file (or paste an SRT with timecodes below)

SRT (with --> timecodes) (used when no audio path)

Words per cue

DaVinci template

Output folder (blank = next to the audio / Downloads)

Render caption video

Render a white-background, word-by-word grey→black caption VIDEO (ASS karaoke + ffmpeg) — N lines × words/line, big readable font, spoken word black / rest grey. Queued; the .mp4 lands in your output folder. Drop it into DaVinci Resolve. No AI, no plugin.

Channel (optional)

Audio file

Font

Words/line

Lines

Size

Output folder (blank = next to audio)

Audio enhance

Clean up a voiceover — local ffmpeg denoise+loudnorm (no upload) / NVIDIA Studio Voice (local RTX GPU) / NVIDIA cloud / Auphonic. Queued; watch the Queue tab.

Channel (optional)

Audio file

Engine

Output folder (blank = next to the input)

Voiceover from script

Load a script file (.md / .txt) and synthesize it to speech (TTS) into your output folder. Queued; watch the Queue tab.

Channel (optional)

Script file

Engine

Voice

Output folder

Script, SEO & Shotlist writer

Generate a full script, SEO metadata, or a shotlist for a topic, then save it into your output folder. Pick the channel for its voice/context. The same generators from Make Video — no project required.

Kind

Channel (voice/context)

Topic / working title

Provider

Key

Model

Extra context / notes (optional)

Output folder

Result (editable — then Save to folder)

Timeline to DaVinci

Assemble a project's media into a DaVinci timeline (FCPXML/OTIO) using a per-channel timeline template — or write a raw FCPXML+OTIO handoff. Preview shows exactly which files will be included + labeled gaps for anything missing. No AI.

Channel (optional)

Project / video folder

Template

Translate video (dub)

Dub a finished video into other languages: transcribe → translate the captions (LLM) → text-to-speech the translation → mux the new audio onto the video. One job per checked language — a translated .srt + a dubbed .mp4 land in your output folder. (Timing is approximate — TTS length vs. the original; use the .srt for fine sync.)

Channel (optional)

Video file

Voice engine (TTS)

Voice

Translate with (LLM)

Key

Model

Output folder (blank = next to the video)

Transcribe

AutoSubs-style speech-to-text on this cockpit's own engine. Local faster-whisper runs on your RTX 3060; outputs a styled .srt + plain .txt into the loaded video's 05_Scripts_and_AI/captions/ (or Downloads).

Source file path

Engine

Channel

Input language

AI transcription model

Translate to English GPU acceleration

Max chars / line

Max lines

Split on punctuation

Text case

Remove punctuation Censor sensitive words Dynamic Time Warping (word timestamps) Speaker labeling

Upscale video (local, GPU)

Source video path

Target

Tool

TTS — text to speech

Text

Engine

Voice

Key

Model

Channel

Save to folder (auto-filled from the loaded video; blank = Downloads root)

Wan (text → video)

Prompt

Engine ⓘ

Key

Model

Channel

Save to folder (auto-filled from the loaded video; blank = Downloads root)

Image generator (thumbnail / any image)

Prompt

Provider

Key

Model

Size

Save directory (blank = Downloads root)

Image editor (qwen-image-edit)

Image URL or path

Provider ⓘ

Key

Model

Strength ⓘ

Channel

Save directory (blank = Downloads root)

Instructions (one per line — applied in order)

HeyGen (avatar A-roll)

Script

Avatar

Voice (name or id)

Channel

Save to folder (auto-filled from the loaded video; blank = Downloads root)

Download B-roll (Pexels + Pixabay)

local downloads Count distinct clips from each service (Pexels + Pixabay) — so Count 3 = 6 clips, 3 per service — and writes an author/attribution .json next to every clip (same name, e.g. pexels_city_1.mp4 → pexels_city_1.json). The saved folder shows in the Queue's Result. online fires the n8n workflow.

Source

Keywords

Save directory

Count / service

Type

Resolution

Competitor research

Niche / topic

Your channel (optional)

Competitor URLs / handles

Provider

Key

Model

Brave search

Perplexity

SearXNG Perplexica

System prompt (edit or leave as-is)

CPM research

Estimate CPM/RPM ranges, revenue projections, and monetization strategies for a niche. Optionally grounded with live Brave web search results.

Niche / topic

Region

Provider

Key

Model

Brave

Perplexity

SearXNG Perplexica

Business plan generator

Generate a concise channel business plan: positioning, content pillars, cadence, monetization roadmap, and 90-day launch plan.

Niche / topic

Provider

Key

Model

Brave

Perplexity

SearXNG Perplexica

Generate subtitles in multiple languages

Transcribes the video once (faster-whisper on your RTX 3060), then translates the captions into every language you tick — writing one <video>.<lang>.srt per language. Each language is a separate job in the Queue (it goes through them one by one). Translation uses the LLM you pick below.

Source media path

Spoken language (source — for accuracy)

Channel

Translate with (LLM)

Key

Model

Output folder (blank = next to the video)

Channel ideation (names / handles / titles)

Kind

Topic / niche

Count

Provider

Key

Model

System prompt (edit or leave as-is)

Channel logo generator

Brand / channel name

Niche / topic

Style

Provider

Key

Model

Size

Count

Save directory (blank = Downloads root)

Checklist generator

Generate a production checklist for a video or channel. Items are stored in Airtable (if configured) or locally. Tick items off as you complete them.

Scope

Ref (channel or video name)

Name	Contact	Email	Niche	Status	Link	Notes
Click Refresh to load sponsors.

Transcript Grabber (YouTube captions)

Pull a video's transcript via the cockpit's transcript connector. Check what tracks are available, fetch the full transcript (segments + char count), then ask for a context block (full / head / summary) to paste into a prompt.

Video URL or ID

Context mode

Deep Search (topic discovery)

Expand a topic/seed into a ranked list of source videos via the deep-search connector. Optionally grab each result's transcript in the same pass — tick the box and pick HOW below.

Topic / seed

Results

grab transcripts

Competitor Monitor (new uploads)

Track competitor channels per channel and check for new uploads since the last run. Add competitors by URL/handle; "Check new" hits the monitor connector and lists fresh uploads.

Channel

Add competitor (URL / @handle / channel ID)

SEO Keyword Analysis

Analyze keyword opportunities, search intent, and title/description angles for a topic. Uses the LLM — pick your provider + model. Output is an actionable markdown report.

Topic / keyword

Provider

Key

Model

Channel

Channel About Generator

Generate a professional YouTube channel About page description. Provide the channel name, niche, and any specifics.

Channel

Niche / topic

Provider

Key

Model

Music Picker (score & SFX)

Pick from your local music/SFX library by brief, or search online sources. Local uses the indexed library; "Index" rebuilds it. Online: Pixabay/YouTube return manual deep-links + search terms.

Brief

Kind

Count

Online sources

Source

Query

Stock B-roll (Pixabay)

Search Pixabay for royalty-free images or videos, preview as a thumbnail grid, and download a pick into your project/Downloads folder.

Query

Media

Save directory (blank = Downloads root)

Poster Annotator (grid overlay + callouts)

Annotate an image using a numbered grid for precise placement. Show the grid to read cell refs (e.g. C4), add annotation rows (arrows / highlights / spotlights / callouts / boxes) anchored to grid refs, then render. Or zoom-crop a single grid cell.

Image path

Annotations

Sec/scene

Zoom crop (grid ref)

Remotion Editor

Export this project's scenes to Remotion, open Remotion Studio to edit directly, or render the master. Prefilled from the active Make Video project — edit the path to point at a different project.

Video / project path

Make Animation (Remotion)

Describe an animation; an LLM turns it into a Remotion scene list (props.json). Open it in Remotion Studio to tweak, and optionally render it now.

Prompt

Save to folder ⓘ

Provider

Key

Model

Size

Render now

Name	ID	Active	Nodes	Actions
Click Refresh (needs N8N_API_KEY in .env).

Pipeline workflows — last run

Open the n8n tab to load recent executions.

All channels (merged Airtable + drive folders). Click a channel to load its projects from All Channels\<ch>\Projects.

Projects

Embed your live Airtable interface URLs here, or read tables natively via /api/airtable/records. Base IDs are in config.yaml / docs/AIRTABLE_MAP.md.

Per-video workspace: pick channel → project → video to see all files + requirements checklist, then run tools on that video.

Describe it → AI builds Remotion scenes → iterate live in the editor tab.

⚡ Quick AI Director

idea or YouTube link → finished MP4 (script · voice · B-roll · SEO)

Idea or reference YouTube link

Scenes

Voice

Visuals

1. Select channel

Channel

2. Select project

Project / Series

3. Select video

Video

Pipeline stages

Each stage is an expandable card: pick its own provider + model (defaulted by the stage's modality), choose Local or n8n, open the stage's tools, tick its checklist, import files, and Run. Default is Local (most n8n workflows are not yet commissioned). Airtable is always mirrored. Each stage's Run enables once that stage has a model chosen (#45); model + n8n-id choices persist per channel.

Export

Ideas from Airtable flagged with Make YouTube Video? toggle. Click Import to load the idea into Make Video. Polls Airtable directly (no n8n needed).

Loading ideas...

Deep production checklist across channel, project, and video scope. Pick a context below; tick items as you go — each toggle saves immediately. Defaults to the active Make Video context if one is loaded.

Checklist context

Channel

Project / Series

Video

Pick a channel (and optionally a project + video) to load its checklist.

Loading checklist…

Scheduled posts/videos from Airtable (Base 2 Social Scheduler, fallback Base 5 Pipeline). Add your Airtable PAT in the API Rotator tab.

Multi-platform scheduler. Real publishing: YouTube uploads via Google OAuth (per channel), and Instagram / X / TikTok / LinkedIn / Facebook publish through your n8n social-publish workflow. The "Connect Account" form just registers a profile so the scheduler knows it exists — it does not hold the real login.

Hootsuite Telemetry: Monitor cost expenditure, local GPU runtimes, API usage metrics, and monthly budget limits. Set custom alerts for budget threshold breaches (both simulated and live usage are tracked).

$0.00

Total Spend (Last 30d)

$0.00

API Cloud Cost

$0.00

Local GPU Amortized Cost

Healthy

Budget Guardrails

API Expenditures by Provider

GPU Consumption Profile

VEED.io local workstation: Run local RTX 3060 speech-to-text, text-based transcript cutting, subtitle burn-ins, and export DaVinci Resolve timelines.

1. Local AI Transcription

2. Hard Subtitle Burn-In

3. FFmpeg Local Utilities

All operations run locally on RTX-3060/CPU. Outputs saved to data/generated/veed/.

📦 Compress

🔄 Convert

🔇 Denoise

📐 Aspect Ratio

🎵 Audio Duck (Sidechain)

vidIQ Analytics & SEO: Optimize titles, generate semantic metadata suggestions, track up to 10 competitor feeds dynamically, and brainstorm AI video concepts.

1. SEO Metadata Optimizer

2. Daily Video Ideas Generator

4. AI Generator Hub

3. Competitor Watch Feed Tracker

Cockpit Channel

InVideo Prompt-to-Video Creator: Autonomously draft AI scripts, fetch royalty-free B-roll stock footage from Pexels, synthesize voiceovers locally, and render completed clips.

1. Prompt to Video

2. Generated Output Player

🎬

No video generated yet

Enter a prompt on the left to start the autonomous pipeline.

Local Asset Library & Brand Kits: Parity with VEED/InVideo workflows. Upload, track, filter, soft-delete local media assets, and manage brand styles locally.
Citations: VEED #21 (Brand Kit Assets) & #22 (Royalty-free Stock Library); InVideo #23 (Unified Brand Kits) & #24 (One-click Brand Overlay).

Create Brand Kit

Your Brand Kits

Loading Brand Kits...

Loading Asset Library...

Which videos performed well vs poorly — YouTube recent videos (sorted by views) + Airtable Base 3. Needs a YouTube key/OAuth and/or an Airtable PAT (add in the API Rotator tab).

Channel

All your API keys, with rotation. Keys from .env are read-only here; keys you add are saved to data\api_keys.json on the server. Add multiple per provider (e.g. several Brave or Alibaba keys) and the cockpit rotates across them. Click fetch to pull a provider's live model list into the Generate dropdowns.

Add a key

Provider

Tag

API key

Note

Local LLM (LM Studio / Ollama)

Point the cockpit at your local model server once; then pick lmstudio or ollama as the Provider in any tool and its Model dropdown fills with the server's models. LM Studio default: http://localhost:1234/v1.

Provider

Base URL

API key (optional)

Default model

enabled

Web Search / Grounding (SearXNG · Perplexica · Perplexity)

Configure the self-hosted search endpoints the research tools can ground with. Enable a source per-tool with its checkbox (Competitor / CPM / Business plan / Sponsors). Perplexity uses a perplexity key (add one in the Keys table below) or runs via your openrouter key. SearXNG hits <base>/search?format=json. Perplexica POSTs <base>/api/search and talks to your LM Studio (OpenAI-compatible) endpoint for its chat/embedding models.

SearXNG base URL

Perplexica base URL

Perplexity default via

Perplexity default model

Perplexica LLM (OpenAI-compatible — your LM Studio)

LM Studio / OpenAI base URL

Key tag (lmstudio)

Chat provider

Chat model

Embedding provider

Embedding model

CLI Agents (Claude Code / Codex / Gemini / Qoder / MiniMax)

Run a local CLI as a "provider" — the cockpit pipes the prompt to the CLI and captures its stdout. No API key: each CLI uses its own login (e.g. claude, codex login, antigravity). Enable one here, then pick it (e.g. antigravity_cli) as the Provider in any tool. Edit the command template if your install differs, and click Test to run it once. The prompt is delivered via prompt via (stdin recommended — never put on the command line).

Loading…

Keys

Provider	Tag	Key	Note	Source	Models	Health	Actions
Loading…

Add / activate / load channels. Shows both Airtable records and existing folders under All Channels\ on the drive. "Add" creates the Airtable record + scaffolds a new folder. Disk-only channels have a Register button to create the Airtable record (no re-scaffold). Add your Airtable PAT in the API Rotator tab.

Settings — master project root

Master project root (base folder that holds All Channels\)

Add a channel

Channel name

Short code

Handle

Channel ID

Niche

Channels active only

Channel	Handle	YouTube	Niche	Source	Active	Folder	Actions
Click Refresh to load channels.

Prompt library for the generation tools (SEO, shotlist, script, ideation…). Set a default per tool, or edit a channel's system prompt. Prompts chosen here appear in the Make Video dialogs.

New prompt

Tool

Label

default for tool

Prompt text

Prompts

Loading…

Editing channel

Channel

All edits below save to this channel only

— pick a channel —

Per-channel system prompt

Copy system prompt from

System prompt for this channel

Per-channel prompts — full set

Every generation tool's prompt for this channel. A channel badge means this channel has its own copy; default means it falls back to the master default. Edit a textarea and click Save to override just that tool (blank reverts to the default). Use Seed all from master to copy every master default into this channel as an editable starting point.

Pick a channel above to load its full prompt set.

Append-only log of every UI click, every cockpit API request, and every outbound provider call (with full request/response). Written to logs/cockpit.jsonl as human-readable plain text. Persists across sessions.

auto-refresh

Click Refresh to load logs.

Select a document to read.

Start / stop / open / view-logs of local server processes you host on this machine — Remotion, SearXNG, ComfyUI, Wan, TTS, Whisper, plus any custom servers you add. Servers launch detached (they keep running if you close the cockpit) and log to data/server_logs/<id>.log. A green dot means something is listening on the server's port.

Add a server

Save a custom server (or override a preset). Use a full command — a list runs without a shell; anything you type here runs through the shell, so docker compose up -d / python main.py work as written. The command is saved to the registry before it can be started.

Name Port Command Working dir URL (optional)

Record screen · camera · PiP · audio

Capture screen, camera, screen+camera (PiP), or pure audio right in the browser via MediaRecorder. Recordings are saved to data/generated/recordings/ on the server.

Mode

Format

Audio?

00:00

Saved recordings

Click Refresh to load.

Teleprompter overlay scrolling script while you record

Load a script and the teleprompter will overlay on your preview while recording. Adjust scroll speed to match your reading pace.

Script text

Scroll speed (50)

Font size

Show teleprompter

Course Maker — one idea → a full course (script, scenes, voice, video)

Course idea

Audience

Modules

Lessons/mod

Voice

Lang

Planner model

Ask AI

Model

🪄 Studio Enhance — AI upscale · background remove · face restore · music gen · scene detect · color grade · visual QA

File path (image or video on an allowed drive)

⬆ Upscale

Scale

Engine

✂ Remove background (image)

😊 Restore faces (image, GFPGAN)

🎵 Generate music

Sec

🎬 Scene detect (video)

🎨 Color grade (video)

Preset

🔎 Visual QA (video, qwen3-vl)

A lightweight chatbot that reuses the cockpit's existing API keys (from the API Rotator) — no extra key needed. Pick any provider you have a key for.

Provider Key Model

Say hello to start a conversation.

Video Quality Coach: reviews a finished render the way a human QC pass would - ffprobe, nine frames from the central 90% with their exact timestamps, a 3x3 contact sheet, then a vision-model verdict against the checklist. It writes only under data/coach/. Accepting a lesson here is what feeds it into the next compile.

Render to review Vision role

idle

Recent reviews

Lesson ledger

Proposed lessons are inert until accepted; rejecting one keeps it on record without feeding generation.

Connections

Key Metrics

Base Summary

Jobs All types video_gen audio_gen image_gen image_edit transcribe upscale render edit publish pipeline n8n All statuses Queued Dispatched Running Done Failed Canceled Refresh Clear finished Clear errors Clear downloads Clear all

Downloads Open Downloads folder Refresh

Channel Profile Defaults

Social Calendar

Resource Telemetry

Teleprompter

Webcam recorder

Audiobook captions

Word-by-word captions

Render caption video

Audio enhance

Voiceover from script

Script, SEO & Shotlist writer

Timeline to DaVinci

Translate video (dub)

Transcribe

Upscale video (local, GPU)

TTS — text to speech

Wan (text → video)

Image generator (thumbnail / any image)

Image editor (qwen-image-edit)

HeyGen (avatar A-roll)

Download B-roll (Pexels + Pixabay)

Competitor research

CPM research

Business plan generator

Generate subtitles in multiple languages

Channel ideation (names / handles / titles)

Channel logo generator

Checklist generator

Sponsors

Transcript Grabber (YouTube captions)

Deep Search (topic discovery)

Competitor Monitor (new uploads)

SEO Keyword Analysis

Channel About Generator

Music Picker (score & SFX)

Stock B-roll (Pixabay)

Poster Annotator (grid overlay + callouts)

Remotion Editor

Make Animation (Remotion)

Pipeline workflows — last run

Recent Executions

Projects

Batch Create Videos

⚡ Quick AI Director

1. Select channel

2. Select project

3. Select video

Generate script

Generate

Voiceover from script

Pipeline stages ⟳ Auto-discover webhook paths

Export

—

Connected Social Accounts

🔌 n8n Publishing Engine

Schedule Multi-Platform Post

Social Queue & Timeline

💾 Saved Resources & Content Scratchpad

Connect Social Account

✨ Live Post Preview Simulator

API Expenditures by Provider

GPU Consumption Profile

Configure Budget Alert

1. Local AI Transcription

2. Hard Subtitle Burn-In

3. Transcript-Based Video Editor 0 Segments

3. FFmpeg Local Utilities

📦 Compress

🔄 Convert

🔇 Denoise

📐 Aspect Ratio

🎵 Audio Duck (Sidechain)

1. SEO Metadata Optimizer

2. Daily Video Ideas Generator

4. AI Generator Hub

Jobs

Downloads

Pipeline stages

3. Competitor Watch Feed Tracker

Channels active only

Saved recordings

Recent reviews

Lesson ledger