Narrative Video Caption Writer (English, Dialogue & Emotion, Scene Understanding)

About OpenTrain

OpenTrain aggregates data-labeling and AI-training jobs from many companies and platforms into a single job board so contributors can discover work without searching dozens of sites. Creating an OpenTrain account is free and applying takes only a few minutes.

We list roles that help train AI systems; this posting connects experienced writers and editors with short-video captioning projects focused on scene understanding and accessibility.

About AI training and this kind of work

AI training (also called data labeling or human feedback) is the human work that teaches models how to interpret and describe content. Tasks include annotating video, transcribing audio, writing concise descriptions, and reviewing model outputs.

This role contributes directly to how multimedia AI systems understand scenes, dialogue, and emotion—improving accessibility and the quality of generated descriptions.

The role

You will produce single narrative captions (50–250 words) for short video clips that prioritize scene understanding: clear setting, background audio, key characters, actions and interactions, overall emotion, and pertinent dialogue. Identify clearly recognizable public figures and brands when visible.

An initial calibration phase precedes ongoing production. Two contributor tracks are available: Captioner and Senior QC, with hourly rates paid per track.

• Captioner rate: $25.75/hr.
• Senior QC rate: $33.00/hr.

What you'll do day to day

Create concise, factual narrative captions for individual clips following detailed style guides and accessibility best practices. Keep captions strictly grounded in on-screen content; do not speculate beyond what is visible or audible.

If you apply to the Senior QC track, you will also edit caption copy, enforce rubrics and style guides, and provide precise written feedback to writers to resolve edge cases and maintain consistency.

• Write 50–250-word captions that: establish setting/background audio, describe key characters and actions, capture overall emotion, include pertinent dialogue, and identify public figures/brands when clearly visible.
• Avoid naming shows or movies, avoid speculation, and avoid over-describing irrelevant background characters.
• Complete an initial calibration test before ongoing assignments.
• Senior QC: review work, apply rubric-based decisions, and deliver constructive written feedback.

Requirements

Applicants must meet the required qualifications below. Senior QC candidates must also meet the additional Senior QC requirements.

Reliable internet capable of streaming source clips is mandatory. All work must be written by you — no AI-assisted writing or machine translation is allowed.

• Required: Native English speaker; USA-based.
• Required: Hands-on experience in one or more of: audio description (DVS), closed captioning/subtitling (SDH), narrative or creative writing, journalism/editorial writing, or script coverage/story analysis for film/TV.
• Senior QC additional: 2+ years in editorial QA / copyediting / caption QA, demonstrated style-guide/rubric enforcement, ability to provide constructive written feedback, and experience resolving edge cases.
• Reliable internet for streaming video clips and participation in calibration.

Nice-to-have skills

These qualifications are optional but relevant and increase your fit for the project and for Senior QC roles.

• Entertainment metadata writing (loglines, synopses, episode/scene summaries).
• Experience in post-production, localization, or accessibility teams.
• Comfort identifying public figures/brands from on-screen context.
• Experience in high-volume, deadline-driven content pipelines (newsrooms, content studios).
• Familiarity with captioning/audio-description best practices and platform accessibility guidelines.
• Comfort using editorial review tools (commenting, change tracking) in collaborative workflows.

How it works and next steps

This is a contractor, part-time opportunity that requires 20+ hours/week. You will apply through OpenTrain, complete a calibration test, and begin production if you meet the quality standards.

Labeling is done in a web-based tool (OTHER). Work is paid hourly by track (Captioner or Senior QC) as listed above. Follow the provided style guide and rubrics closely; calibration ensures your output matches project standards before ongoing assignments begin.

• Employment type: Contractor, Part-time.
• Time requirement: 20+ hours per week.
• Data type: Video; label types: Text generation and text summarization.