AICreatorsHow-To

Get Paid to Share Photos and Clips: A Beginner’s Checklist for Submitting to AI Training Marketplaces

ffreestuff

2026-02-03

10 min read

Step-by-step checklist and email templates to submit photos and vertical clips to Human Native–style AI marketplaces and Holywater scouts in 2026.

Stop wasting time on expired offers — get paid to submit your best photos and short clips to AI marketplaces and scouts

If you shoot photos or short-form vertical video and you want real pay for your work, this checklist walks you through everything to submit content the right way to AI-training marketplaces like Human Native and content scouts at platforms such as Holywater. In 2026 the market is moving fast: companies are buying creator-owned training data and streaming platforms are paying for vertical-first clips. Use this guide to avoid wasted submissions, protect your rights, and get paid.

Why 2026 is the moment to pitch AI marketplaces and vertical-video scouts

Two developments made 2025→2026 a turning point for creators:

Marketplace payouts for training data: In January 2026 Cloudflare acquired Human Native, accelerating demand for creator-supplied training data where developers pay for real-world content. This means structured marketplaces are now a viable revenue stream for photographers and videographers.
Vertical video scale: Platforms like Holywater raised fresh funding in early 2026 to scale mobile-first episodic and microdrama content. Scouts need short, high-quality vertical-first clips and serialized frames that perform on phones.

These shifts create two distinct buyer profiles: AI research/enterprise buyers (training datasets, labeled images/clips) and entertainment/streaming buyers (short-form vertical assets). Your submission must match the buyer’s needs.

Quick overview: What this guide gives you

A pre-submission checklist so your files aren’t rejected
Technical specs for both photos and video clips
A ready-to-send creator pitch / submission email template for Human Native–style marketplaces and for Holywater scouts
Negotiation tips for licensing & pricing, payment, and contracts

Before you submit: the must-have checklist

Submit with confidence. This checklist prevents common rejections and speeds payouts.

Rights & Releases
- Model releases for any identifiable person (signed, dated PDFs).
- Property releases for private property or trademarked elements.
- Confirm you own the copyright or have assignment/clear permission to sell.
Licensing clarity
- Decide non-exclusive vs exclusive (non-exclusive is easier to place).
- Determine permitted uses: AI training, commercial use, redistribution — list them.
Metadata & Provenance
- Embed EXIF and IPTC where possible. Provide a sidecar JSON/CSV manifest for datasets.
- Include creation date, location (if allowed), camera settings, and keywords.
Tech hygiene
- Clean edits — remove watermarks, dust, sensor spots.
- Supply checksums (MD5 or SHA256) for large transfers and S3 presigned uploads.
File packaging
- Use clear filenames: creatorname_assettype_YYYYMMDD_sequence.ext (example: jdoe_img_20260112_001.CR2).
- Compress non-lossy where possible: ZIP or TAR.GZ for batches; provide thumbnail JPGs.
Payment & tax setup
- Have PayPal, Stripe, or bank details ready; know required tax forms (W-9, W-8BEN for U.S. platforms).

Technical specs: photo and video minimums buyers ask for in 2026

Specifications differ between AI-training and streaming buyers. Below are practical, platform-ready specs.

Photos (recommended baseline)

Preferred formats: RAW (CR2/NEF/ARW) plus high-res JPEG for previews (minimum 4,000 px on longest side).
Color: sRGB for web delivery; provide ProPhoto or Adobe RGB for premium buyers on request.
Compression: Lossless for datasets; 12–16-bit where possible for scientific/AI uses.
Metadata: EXIF (camera/ISO), IPTC (caption, creator, contact, release info), and JSON sidecar with taxonomy and tags.
Variety: For AI training, submit diverse examples (angles, lighting, devices, skin tones, ages, contexts).

Short-form video clips (the Holywater & vertical-first playbook)

Aspect ratio: 9:16 vertical (1080x1920 minimum). Provide 4:5 and 1:1 crops if possible.
Resolution: 1080p is common; 4K vertical is a premium sell (2160x3840).
Frame rate: native fps (most buyers accept 24/25/30/60); mention variable frame rate sources.
Codecs: H.264 for quick preview; deliver original ProRes/HEVC/Apple ProRes for final licensing.
Duration: 5–60 seconds for microcontent; supply loopable 6–15s versions for ads/UGC-style uses.
Audio: Provide dry and mixed audio tracks; 48 kHz WAV preferred for masters.
Captions & transcripts: Include SRT/WEBVTT and a plain-text transcript for accessibility and search.

Metadata & manifest: the single most under-used trick

For dataset buyers, a simple manifest speeds acceptance. Use a CSV or JSON with one line per asset including:

filename, checksum, format, resolution, duration (video), tags, location (if allowed), model_release (Y/N), property_release (Y/N), contributor, contact_email, suggested_license, suggested_price

Example CSV header: filename,sha256,format,width,height,duration,tags,model_release,property_release,contributor,email,price_usd

Submission approach: Human Native–style marketplaces vs Holywater scouts

Match your pitch to the buyer’s intent.

AI-training marketplaces (Human Native-style)

Buyers value diversity, clear licensing for AI training data, and labeled metadata.
Prioritize RAW or high-bit-depth images, lots of labeled variants, and robust manifests.
Make usage explicit: “Allowed for model training, embeddings extraction, and commercial redistribution.”

Vertical streaming & content scouts (Holywater-style)

Buyers want final-look verticals: strong storytelling hooks, jump cuts, captions, and loopable clips.
Provide short cuts, teasers, and localized caption files. Include talent contact if performance rights require it.

Pricing & licensing: how to ask without scaring buyers off

There’s no single rate. Use market norms to guide your ask.

Non-exclusive AI training license: baseline per-image or per-clip micro-license (example: $2–$25 per image for bulk datasets; clips often $10–$100 depending on uniqueness).
Exclusive use or premium vertical placements: ask higher: $200–$5,000 depending on exclusivity term, territory, and distribution.
Royalty vs one-time: For AI, marketplaces often prefer one-time dataset buys. For streaming, negotiate revenue share or CPM-style royalties where feasible.

Tip: provide a pricing band and leave room to negotiate. Be explicit about exclusivity term and purpose. Example clause: “Non-exclusive, worldwide license for model training and derivative model outputs; 12-month exclusivity negotiable for +X fee.”

Submission workflow: step-by-step

Prepare package: thumbnails, master files, manifest, releases, pricing sheet.
Upload: S3 presigned URL preferred for large batches; otherwise secure cloud link with password-protected ZIP.
Send pitch email: concise subject, bullet list of assets, link to manifest, checksum, and contact details.
Follow up: 5–7 business days is typical. If you don’t hear back, send a polite reminder with a one-click download link for the content.
Negotiate & sign: request a short PO or simple license agreement. Don’t accept vague “rights to use” terms—get purpose and duration in writing.
Deliver & invoice: include invoice number and payment terms (Net 15/30). Attach tax form if requested.

Submission email templates — copy, paste, personalize

Below are two tight templates. Replace placeholders with your details. Keep the subject lines crisp.

Template A: Pitch to a Human Native–style AI training marketplace

Subject: Submission: 1,200 labeled images + 150 clips (AI training) — [YourName]

Hi [Marketplace Contact Name],

I’m [Full Name], a photographer/videographer based in [City/Country]. I’m submitting a package of images and short clips suitable for AI training. Quick summary:

• Assets: 1,200 images (RAW + JPG), 150 short clips (vertical + horizontal) 
• Themes/tags: street scenes, indoor dining, multi-ethnic portraits, product close-ups
• Releases: 200 model releases (PDF), 30 property releases
• Metadata: CSV manifest included (filename, sha256, tags, model_release)
• Delivery: S3 presigned URL (link) — all files zipped; checksums attached

Suggested license: Non-exclusive, worldwide, for model training, embeddings, and derivative use. Suggested one-time price: $[range]. Open to negotiation for volume or exclusivity.

Attached: sample preview sheet (50 JPGs), CSV manifest, sample model release. Full package available at the link above.

Happy to hop on a 15-min call to discuss labeling or custom annotation needs.

Thanks,
[Full Name]
[Website link / portfolio]
[Phone]
[Payment details & tax ID if already set up]

Template B: Pitch to a Holywater-style scout (short vertical video)

Subject: Vertical clip submission — 40 x 9:16 drama-ready microclips — [YourName]

Hi [Scout Name],

I’m [Full Name], I create short-form vertical content optimized for mobile-first platforms. I’m sending a pack of 40 clips (6–30s) that match Holywater’s microdrama needs:

• Format: 9:16 vertical, 1080x1920 (4K masters available)
• Content: micro-moments, emotional beats, 4 loopable 8s clips per scene
• Captions/Subs: SRTs included; dry audio tracks available
• Talent: cast listed with signed releases

Delivery: password-protected WeTransfer (link) — preview thumbnails attached.

License preference: non-exclusive short-form streaming license. Asking $[X] per clip or $[bundle price] for exclusive 3-month window. Open to trial placements.

Portfolio: [link]. Can deliver fresh custom shoots on 7–10 day turnaround.

Best,
[Full Name]
[IG/portfolio]
[Phone]

What to expect after you hit send

Initial triage: you’ll get an auto-reply or human confirmation within 48–72 hours at professional marketplaces.
Technical review: they check manifests, releases, and checksums — expect requests for missing model releases.
Commercial review: price and licensing negotiation; some platforms ask for exclusivity windows.
Contract & NDA: expect a short license agreement; for enterprise buyers you may receive a standard NDA before detailed briefings.
Payment & delivery: once signed, deliver masters; payments typically 15–60 days depending on platform terms.

Common rejection reasons and how to avoid them

No model release for an identifiable person. Fix: collect releases or blur faces for training use.
Poor or missing metadata. Fix: provide a manifest with tags and captions.
Wrong aspect ratio or codec. Fix: include platform-preferred preview and master files.
Unclear licensing language. Fix: state exact permitted uses and whether training is included.

Advanced strategies & 2026 predictions for creators

Position yourself for recurring revenue:

Build task-specific packs: curated bundles for object detection, facial emotion, or mobile UGC—buyers prefer ready-to-use sets.
Offer labeling services: creators who provide annotated assets (segmentation, bounding boxes, transcripts) command higher rates. See practical data patterns that reduce cleanup costs here.
API-first delivery: marketplaces are adopting programmatic ingestion (presigned S3, API endpoints). Automate packaging with a script that generates manifests and checksums.
Track provenance: embed stable IDs and supply chain metadata; expect buyers to ask for provenance to meet AI governance rules (e.g., EU transparency requirements rolled out in 2025–2026).
Repurpose vertically: reshoot or re-edit horizontal footage into high-value verticals for Holywater-style placements.

Market prediction: by late 2026, more centralized marketplaces will offer tiered payouts for labeled vs unlabeled data. Early movers who supply labeled, diverse, and well-documented assets will earn the most.

Quick reality check: marketplaces are professional buyers. Treat submissions like B2B sales — neat packaging, clear licensing, and predictable delivery win contracts.

Mini-case study (illustrative)

A creator who packaged 2,000 labeled street-photos into 10 themed bundles and offered S3 delivery and CSV manifests closed a non-exclusive dataset sale in 2025 after a 2-week review. The deal included a small bulk bonus for high diversity in lighting and age groups. The success factor: metadata + releases + modular pricing.

Checklist recap: 10-point pre-submission (printable)

Collect model/property releases (PDFs)
Decide license type & exclusivity terms
Export master files + preview JPGs/MP4s
Create manifest (CSV/JSON) with checksums
Include SRT/transcripts for clips
Prepare pricing band and negotiation notes
Provide contact & payment/tax info
Upload to secure cloud (S3 presigned preferred)
Send concise submission email using the templates above
Follow-up after 5–7 business days

Final checklist tips to protect yourself

Never sign away “all rights” without payment tied to exclusivity and term limits.
Keep copies of original files and release forms in a secure cloud folder.
Use simple contract language and request clarifying examples of intended downstream uses.

Take action now — 3-step plan for your next submission

Pick 50–200 high-quality assets. Prepare releases and a manifest.
Choose the right buyer (Human Native-style for datasets; Holywater-style for verticals) and tailor the template email above.
Send the pitch, follow up in one week, and be ready to negotiate licensing terms.

If you want a one-page downloadable manifest template or a customizable submission email in DOCX, click through to our free creator pack — tested for the Human Native and Holywater workflows. Start small, be organized, and treat each submission like a product pitch. The market pays creators who bring clarity, compliance, and ready-to-use assets.

Call to action

Use the checklist above and pick one pack to submit this week. Want the editable email templates and manifest CSV? Sign up for our creator pack and get a free template bundle plus a weekly roundup of verified AI training and vertical-video buyer opportunities.

freestuff

Contributor

Senior editor and content strategist. Writing about technology, design, and the future of digital media. Follow along for deep dives into the industry's moving parts.