Local AI Video Pipeline Developer – ComfyUI, n8n, Python & FFmpeg

Please login or register as jobseeker to apply for this job.

: TYPE OF WORK

Gig

: SALARY

8–$12/hour

: HOURS PER WEEK

20

: DATE UPDATED

Aug 4, 2026

JOB OVERVIEW

LOCAL AI VIDEO PIPELINE DEVELOPER
ComfyUI, n8n, Python, FFmpeg, Windows 11, and NVIDIA GPU

ABOUT THE PROJECT

Wealth Crafted Futures LLC is building a completely local, automated AI short-video production system.

The finished system must allow a non-technical user to place a complete story or script into an input folder, select a few project options, and start production without manually operating multiple AI programs.

This is not a normal video-editing position. I need an experienced developer and systems integrator who can connect reliable existing open-source programs, repositories, and models into one working local pipeline.

The target computer runs Windows 11 with:

• NVIDIA RTX 5080 16 GB
• AMD Ryzen 9 5950X
• 96 GB RAM
• Multiple NVMe drives
• No paid cloud services or paid AI APIs

PRIMARY RESULT

I should be able to:

1. Add a complete story or script.
2. Select the desired language, voice mode, and production settings.
3. Click Start.
4. Have the system automatically create the project folders.
5. Break the story into episodes, scenes, and shots.
6. Generate the video segments.
7. Generate narration or individual character voices.
8. Generate accurately timed captions.
9. Add audio and assemble the segments.
10. Export completed vertical short-form videos.
11. Resume a job after an interruption without restarting completed work.

PIPELINE REQUIREMENTS

The finished pipeline must:

• Run locally on Windows 11.
• Be optimized for an NVIDIA RTX 5080 with 16 GB VRAM.
• Use only free and open-source local software.
• Have zero required paid APIs, cloud services, subscriptions,s or usage fees.
• Use existing reliable open-source repositories before building features from scratch.
• Verify that every selected repository works on Windows and the RTX 5080.
• Verify the license and commercial-use requirements of every dependency.
• Include n8n or another suitable local workflow orchestrator.
• Use Python for custom automation and integration where needed.
• Use FFmpeg for video and audio processing.
• Use ComfyUI or another proven local generative-video workflow system.
• Provide a simple local web interface or desktop interface for the user.
• Include logging, retries, checkpoints, and clear error messages.

PROJECT-FOLDER AUTOMATION

When a newscript is added, the system must automatically create a new project folder named after the story.

The project folder must contain organized subfolders for:

• Original script
• Story bible
• Character registry
• Character reference images
• Locations
• Props
• Scene breakdowns
• Shot plans
• Prompts
• Continuity information
• Generated images
• Generated video segments
• Narration and character voices
• Music and sound effects
• Captions
• Intermediate renders
• Completed episodes
• Quality-control reports
• Logs
• Project configuration
• Recovery checkpoints

STORAGE RULES

All permanent programs, repositories, models, project data, and completed outputs must be stored on the D: drive.

Temporary runtime processing must use:

E:\WCF_Short_Video_Temp

The E: drive is for active-job scratch data, extracted frames, temporary renders, caches, interpolation files, upscaling files and other transient processing.

Completed or required files must be moved back to the appropriate D: drive project folder.

VIDEO STRUCTURE

Each completed episode must be approximately one minute long.

Each one-minute episode will normally be created from four connected 15-second video segments.

The system must preserve:

• Character appearance
• Clothing
• Locations
• Props
• Story details
• Camera direction
• Voice assignment
• Visual style
• Continuity between segments
• Continuity between episodes

A normal series will contain 20 connected one-minute episodes with a real beginning, middle and ending.

The pipeline must also allow a series to be extended beyond 20 episodes when selected.

VOICE, LANGUAGE AND CAPTION OPTIONS

The user must be able to select:

• Narrator-only mode
• Individual character voices
• Narrator plus character voices
• Output language
• Caption language
• Burned-in captions
• Separate SRT or VTT caption files

Narration, character voices,s and captions must be generated locally.

The system should support adding more languages later without rebuilding the entire pipeline.

SCHEDULING AND JOB CONTROL

The pipeline must support a production schedule that can run from approximately 8:00 AM until 6:30 AM the following day, or stop after producing 20 completed one-minute episodes.

It must include:

• Start
• Pause
• Resume
• Stop after current step
• Retry failed step
• Skip an approved failed item
• Production queue
• Completed-job history
• Error log
• GPU-memory error handling
• Safe recovery after Windows restarts

USER INTERFACE

The user should not need to edit code or manually operate every underlying program.

The interface should include:

• Select or import script
• Project name
• Visual style
• Aspect ratio
• Resolution
• Number of episodes
• Narrator or character-voice mode
• Language selection
• Caption options
• Start, pause, and resume controls
• Current production stage
• Progress for each episode
• Error and warning display
• Preview and approval controls
• Open project folder button

DELIVERABLES

The developer must provide:

• Complete source code
• Git repository with organized commits
• All n8n workflows
• All ComfyUI workflows
• Python scripts
• Configuration files
• Dependency and model list
• License report for every major dependency
• Windows installation scripts
• Start and shutdown scripts
• Backup and recovery instructions
• User manual
• Developer documentation
• System architecture diagram
• Troubleshooting guide
• Test scripts
• Example project
• No hidden paid services or personal developer accounts

The system must not depend permanently on the developer’s computer, account, API key, cloud storage, or subscription.

MILESTONE DOCUMENTATION

At every stopping point or milestone, the developer must create or update a self-contained Markdown review file named PROJECT_REVIEW.md.

It must contain:

• Exact repository state
• Work completed
• Files created or changed
• Test results
• Failed tests
• Problems discovered
• Fixes applied
• Known limitations
• Current Git branch and commit
• Any authorization needed
• Exact recommended next action

ACCEPTANCE TEST FOR THE FIRST WORKING VERSION

The first working version will be accepted when it can:

1. Accept a sample script through the user interface or input folder.
2. Create the complete story project folder structure.
3. Break the script into scenes and four connected shots.
4. Generate or process four approximately 15-second segments.
5. Generate local narration or character voices.
6. Generate synchronized captions.
7. Join the four segments into one completed one-minute vertical MP4.
8. Save all permanent project files to the D: drive.
9. Use the E: drive for temporary processing.
10. Resume after an intentional interruption.
11. Produce logs explaining every completed or failed step.
12. Complete the process without a paid API or cloud-generation service.

REQUIRED EXPERIENCE

Applicants should have real hands-on experience with several of the following:

• Python
• n8n self-hosting and workflow development
• ComfyUI workflows and custom nodes
• FFmpeg
• Windows automation
• NVIDIA CUDA and VRAM optimization
• REST APIs and local APIs
• Git and GitHub
• Docker where appropriate
• Local text-to-speech
• Local transcription and caption generation
• Generative image or video models
• Queue systems, checkpoints and error recovery

HOW TO APPLY

Start your application with the exact words:

LOCAL-5080

Applications that do not begin with LOCAL-5080 may not be reviewed.

Please answer all of these questions:

1. Describe the most similar local AI or video-automation system you have personally built.
2. Provide a link, screenshot, or short screen recording showing a real ComfyUI, n8n, or Python automation you built.
3. Have you installed and optimized local AI software on Windows with an NVIDIA GPU?
4. What open-source tools would you initially evaluate for this project?
5. How would you prevent character and story continuity from being lost between video segments?
6. How would you make a failed production job resume without repeating completed work?
7. How would you verify software licenses and prevent required ongoing fees?
8. What would you build during the first paid milestone?
9. Estimate the hours required for the proof of concept and the first usable MVP.
10. State your hourly rate and weekly availability.

Do not send only a generic AI-generated cover letter. Specific technical answers and proof of previous work are required.

HIRING PROCESS

The hiring process will include:

• Portfolio and technical-answer review
• Short video interview
• Paid technical planning task
• Paid proof-of-concept milestone
• Full build divided into approved milestones

No applicant will be asked to complete custom production work for free.

This begins as a project-based position. Continued maintenance and future automation projects may be available for someone who produces reliable, documented work.

SKILL REQUIREMENT

Video Editing Python AI Automation

SHARE THIS POST

Job Post	Reporter	Rank	Reason	Date

BENCHMARKS

Loading Time: Base Classes	0.0008
Controller Execution Time ( Jobseekers / Job )	0.0110
Total Execution Time	0.0123

GET DATA

No GET data exists

MEMORY USAGE

1,542,024 bytes

POST DATA

No POST data exists

URI STRING

jobseekers/job/Local-AI-Video-Pipeline-Developer-ComfyUI-n8n-Python-FFmpeg-1702588

CLASS/METHOD

jobseekers/job

DATABASE: onlinejobs (Jobseekers:$db) QUERIES: 13 (0.0061 seconds) (Hide)

0.0003	`SELECT * FROM exrates WHERE rate_name = 'USD-PHP'`
0.0003	SELECT * FROM `employer_jobs` WHERE `job_id` = 1702588 LIMIT 1
0.0003	SELECT * FROM `employers` WHERE `employer_id` = 926216 LIMIT 1
0.0007	SELECT COUNT(*) AS `numrows` FROM `t_thread` `t` LEFT JOIN `t_thread_misc` `misc` ON `t`.`id` = `misc`.`thread_id` WHERE `t`.`job_id` = 1702588 AND `misc`.`id` IS NULL
0.0004	SELECT e.business_name, e.logo, e.website, e.rebill_date, e.date_added member_date, hits, DATEDIFF('2026-08-09',ej.date_added) duration_days, DATEDIFF('2026-08-09',e.rebill_date) duration_rebill, ej.*, e.deactivate FROM employers e, employer_jobs ej WHERE e.employer_id = ej.employer_id AND ((e.user_level >= '500' AND ej.date_added <= e.rebill_date) OR e.employer_id = '' OR (ej.date_approved <> '2000-01-01' and DATEDIFF('2026-08-09',ej.date_added) <= 14 )) AND e.deactivate != 1 AND ej.deleted = 0 AND job_id = '1702588'
0.0002	SELECT * FROM `employer_jobs_skills` `ejs` LEFT JOIN `skills_categories` `sc` ON `ejs`.`skill_id` = `sc`.`id` WHERE `job_id` = 1702588
0.0013	`UPDATE employer_jobs SET hit_counts = '*Aug-04-2026=416Aug-05-2026=164Aug-06-2026=97Aug-07-2026=36Aug-08-2026=28*Aug-09-2026=1' WHERE job_id= '1702588'`
0.0005	`UPDATE employer_jobs SET monthly_hits = '***Aug-2026=742' WHERE job_id= '1702588'`
0.0007	`SELECT date_sent FROM jobseeker_sent_emails WHERE jobseeker_id = '' AND job_id = '1702588' AND status LIKE 'sent%' ORDER BY id DESC`
0.0002	SELECT * FROM `employer_jobs_skills` `ejs` LEFT JOIN `skills_categories` `sc` ON `ejs`.`skill_id` = `sc`.`id` WHERE `job_id` = 1702588
0.0007	SELECT COUNT(*) AS `numrows` FROM `employer_jobs` WHERE `employer_id` = '926216' AND `date_added` >= '2022-06-08'
0.0003	`select * from teasers`
0.0002	`SELECT * FROM skill_categories WHERE skill_cat_id=''`

HTTP HEADERS (Show)

SESSION DATA (Show)

__ci_last_regenerate	1786300159
last_page	https://v1.stage.onlinejobs.ph/jobseekers/job/Local-AI-Video-Pipeline-Developer-ComfyUI-n8n-Python-FFmpeg-1702588
csrf-token	24492a065dd53cfc2c286fff2cbd57b7

CONFIG VARIABLES (Show)

base_url	https://v1.stage.onlinejobs.ph
log_threshold	2
enable_profiler
index_page
uri_protocol	PATH_INFO
url_suffix
language	english
charset	UTF-8
enable_hooks	1
subclass_prefix	MY_
composer_autoload	vendor/autoload.php
permitted_uri_chars	a-z 0-9~%.:_\=\-\+\@?\&
enable_query_strings
controller_trigger	c
function_trigger	m
directory_trigger	d
allow_get_array	1
log_path
log_file_extension
log_file_permissions	420
log_date_format	Y-m-d H:i:s
error_views_path
cache_path
cache_query_string
encryption_key	OdBUArjiWg9I7u7bvAwQ7Fu35VB1kzga
sess_driver	files
sess_cookie_name	ci_session
sess_expiration	2678400
sess_save_path
sess_match_ip
sess_time_to_update	300
sess_regenerate_destroy
cookie_prefix
cookie_domain
cookie_path	/
cookie_secure
cookie_httponly
standardize_newlines
global_xss_filtering	1
csrf_protection
csrf_token_name	csrf_test_name
csrf_cookie_name	csrf_cookie_name
csrf_expire	7200
csrf_regenerate	1
csrf_exclude_uris	Array ( )
compress_output
time_reference	local
rewrite_short_tags
proxy_ips
ojadmin	JpjZkQN5A8l@^L
salt	+UFjSAT49tPZLtmU2CIG2FYN7pRhgsWyLHgSyQa6k3I=
queue_enabled
twilio	Array ( [sid] => AC00c0594045c6eef9407e8fff01f3d467 [token] => f0bfc0b73444a077894a43f5f75e6d41 [length] => 6 [from] => +639221200200 [code_expiry_seconds] => 14400 )
maintenance	Array ( [admin] => Array ( [verification] => ) [announcement_bar] => Array ( [show_slow_issue_message] => ) )
trustpilot	Array ( [to_email] => onlinejobs.ph+3d844ebf71@invite.trustpilot.com )
services	Array ( [google_tag_manager] => Array ( [id] => GTM-T5CQMS6P ) [chatgpt] => Array ( [secret] => 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 ) [claude] => Array ( [secret] => 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 ) )
v2_url	https://v2.onlinejobs.ph
replacemyself_secret_key	Kk1WpgTMkc4wBQOqC5rqCssdLhACKrsJeTtO1ywUkT4=
app	Array ( [app] => Array ( [command] => Array ( [map] => Array ( [OnlineJobs\Mailer\Command\SendEmailCommand] => OnlineJobs\Mailer\Handler\SendEmailHandler [OnlineJobs\Mail\Command\SendEmailCommand] => OnlineJobs\Mail\Handler\SendEmailHandler [OnlineJobs\Mail\Command\QueueEmailPendingCommand] => OnlineJobs\Mail\Handler\QueueEmailPendingHandler [OnlineJobs\Mail\Command\QueueJobEmailPendingCommand] => OnlineJobs\Mail\Handler\QueueJobEmailPendingHandler [OnlineJobs\Mail\Command\QueueUnapprovedEmailCommand] => OnlineJobs\Mail\Handler\QueueUnapprovedEmailHandler [OnlineJobs\Mail\Command\ProcessEmailQueueCommand] => OnlineJobs\Mail\Handler\ProcessEmailQueueHandler [OnlineJobs\Mail\Command\PutJobseekerEmailOnHoldCommand] => OnlineJobs\Mail\Handler\PutJobseekerEmailOnHoldHandler [OnlineJobs\Purchase\Command\RealEstateVaCourseCommand] => OnlineJobs\Purchase\Handler\RealEstateVaCourseHandler [OnlineJobs\Test\Command\HelloCommand] => OnlineJobs\Test\Handler\HelloHandler ) [middleware] => Array ( ) ) [providers] => Array ( [0] => OnlineJobs\Bus\BusServiceProvider [1] => OnlineJobs\Database\DbServiceProvider [2] => OnlineJobs\Mailer\MailerServiceProvider [3] => OnlineJobs\Support\SupportServiceProvider [4] => OnlineJobs\Employer\ServiceProvider [5] => OnlineJobs\Jobseeker\ServiceProvider [6] => OnlineJobs\Mail\ServiceProvider [7] => OnlineJobs\Postal\ServiceProvider [8] => OnlineJobs\Queue\ServiceProvider [9] => OnlineJobs\Courier\ServiceProvider [10] => OnlineJobs\Job\ServiceProvider [11] => OnlineJobs\Purchase\ServiceProvider ) ) )
meta	Array ( [facebook] => Array ( [input] => 1 [connect] => 2 [timezone] => 3 [url] => 30 [id] => 31 [photo] => 32 [name] => 33 [email] => 34 [review] => 35 ) [trust] => Array ( [government_id] => 4 [utility_bill] => 5 [selfie_photo] => 6 [profile_picture] => 7 [phone_number] => 8 [facebook] => 9 ) [verify] => Array ( [profile_picture] => 10 [government_id] => 11 [selfie_photo] => 12 [utility_bill] => 13 [address] => 14 [name] => 15 [reviewed] => 16 [government_group] => 28 [address_group] => 29 [mobile] => 38 ) [address] => Array ( [complete] => 17 [room_unit_no] => 20 [house_no] => 21 [street] => 22 [subdivision] => 23 [barangay] => 24 [city] => 25 [province] => 26 [postcode] => 27 ) [reviewed] => Array ( [similar_names] => 1 ) [name] => 18 [phone] => 19 [sms_verification_code] => 36 [resend_verification_code] => 37 [address_read] => 39 [first_name] => 100 [middle_name] => 101 [last_name] => 102 [edit] => Array ( [government_id_group] => 40 [address_group] => 41 [mobile_group] => 42 ) [salary_re_entry] => 103 [id_proof_recalibrate] => 104 [email] => 44 [mobile_verification_type] => 45 )
honeypot	Array ( [name_field] => my_name [enabled] => 1 )

HTTP_ACCEPT	/
HTTP_USER_AGENT	Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)
HTTP_CONNECTION	Upgrade
SERVER_PORT	80
SERVER_NAME	v1.stage.onlinejobs.ph
REMOTE_ADDR	127.0.0.1
SERVER_SOFTWARE	Apache/2.4.58 (Ubuntu)
HTTP_ACCEPT_LANGUAGE
SCRIPT_NAME	/index.php
REQUEST_METHOD	GET
HTTP_HOST
REMOTE_HOST
CONTENT_TYPE
SERVER_PROTOCOL	HTTP/1.1
QUERY_STRING
HTTP_ACCEPT_ENCODING	gzip, br, zstd, deflate
HTTP_X_FORWARDED_FOR	216.73.216.227
HTTP_DNT

Local AI Video Pipeline Developer – ComfyUI, n8n, Python & FFmpeg

Please login or register as jobseeker to apply for this job.

TYPE OF WORK

SALARY

HOURS PER WEEK

DATE UPDATED

Why is this blurred?