How to Pitch Your Catalog to AI Video Startups: A Composer’s Outreach Template

Hook: Turn AI video demand into reliable income — without getting steamrolled

As AI-specific risks ( model training, unlimited derivatives, and perpetual sublicensing) make outreach feel like stepping into negotiations without a parachute. This guide gives you a battle-tested outreach playbook, ready-to-send scripts, practical pricing models, and negotiation tactics tailored for companies like Higgsfield and Holywater in 2026.

Topline: What to expect in 2026 marketplace dynamics

Short version: well-funded AI video firms are buying or licensing music aggressively to train models, fuel creator toolkits, and clear short-format syncs for vertical video. Higgsfield’s rapid scale and Holywater’s new funding round (late 2025–early 2026) mean a growing pool of buyers — but also competition from large catalogs and data marketplaces. Expect four dominant deal types:

• Traditional sync license for specific uses (ads, episodes, clips).
• Dataset / training license allowing model training, often with different financial and legal structures.
• Subscription or per-use micro-sync integrated into creator UIs for social videos.
• Revenue-share integrations where you get a cut of in-app monetization.

Recent industry moves (Cloudflare’s Human Native purchase in early 2026, Higgsfield’s valuation and Holywater’s funding) show momentum toward creator-paid data systems and marketplaces, so composers can convert one-off syncs into recurring revenue with the right terms.

How to approach AI video startups — 6-step outreach framework

1. Research: Map the company’s product lines (creator tools, studio features, automated editors). For Higgsfield look at short-form creator workflows; for Holywater focus on episodic vertical content.
2. Segment your catalog: Create three buckets — hot (short, loop-friendly hooks under 30s), cinematic (30–180s), and dataset (isolated stems, metadata (BPM, key, instrument tags), MIDI, dry mixes).
3. Prep deliverables: Provide WAVs (44.1/48kHz), stem packs, high-res metadata (BPM, key, instrument tags), and usage notes. Attach a simple licensing one-pager.
4. Craft an outreach pitch: Keep it short, quantify value (track count, genres, stem availability), and suggest deal frameworks.
5. Negotiate: Lead with non-exclusive offers and test campaigns, then scale into exclusivity or dataset licensing with better economics.
6. Close & scale: Add reporting, timestamps for usage, and a renewal cadence.

Outreach scripts you can copy and send

Below are concise, customizable messages for email, LinkedIn, and follow-ups. Keep your first outreach under 120 words.

Email: Cold pitch (first touch)

Subject: Catalog licensing — 120 tracks of short-form hooks + stems for Higgsfield/Holywater

Hi [Name], I’m [Your Name], composer/producer with a 250-track catalog focused on short-form hooks, ambient beds, and stems optimized for vertical video. I’ve packaged ready-to-integrate WAVs + stems and detailed metadata (BPM, key, loop points). I’d love to discuss a pilot non-exclusive sync + dataset license for your creator tools or episodic content. Typical pilot asks: 20 tracks, 3 months, $X upfront + royalty. Attached: one-pager + 5 demo links. Available for a quick 15-minute call next week? Thanks — [Name] / [Phone] / [Link to portfolio]

LinkedIn: Short message for product/partnership lead

Hi [Name], I compose catalog-ready hooks and stems tailored for mobile verticals. I have a 20-track pilot package and a dataset-ready stem pack — would you be open to a short pilot license conversation this week? — [Name]

Follow-up (3 business days)

Hi [Name], quick follow on my note — I can prepare a 20-track pilot pack with metadata and 5x stems per track by EOD. Typical pilot structure I recommend: $X upfront, 3-month non-exclusive, track-level reporting. Happy to tailor. — [Name]

Phone / Zoom opener (first 90 seconds)

Thanks for taking the call. Quick context: I license music to creators and platforms. I’ve prepared a pilot pack (20 short hooks + stems) that integrates into TikTok-style editors and episodic packs. I’m proposing a 3-month non-exclusive test — $X upfront for full stems and usage reporting, plus a modest revenue share if your creators monetize. My goal is a low-friction pilot to prove engagement. How do you typically evaluate new catalog partners?

Pricing models — practical numbers and examples (2026 market context)

Pricing depends on the buyer, exclusivity, and rights. Use these 2026-informed models as starting points; tailor by track quality, catalog size, and the buyer’s stage.

1) Sync license (short-form/creator tool) — non-exclusive pilot

• Structure: Per-track upfront + per-use micro-fee or CPM-like royalty.
• Example pricing: $150–$500 per 30–60s track upfront for non-exclusive short-form hooks; add $0.002–$0.01 per stream/view or 2–5% share of in-app creator revenue.
• Why: Low barrier for platforms, recurring upside via micro-usage.

2) Dataset / training license (most sensitive)

• Structure: Upfront dataset fee + tiered per-model usage or revenue share + strict usage limits (no resale of raw files, no public redistribution).
• Example pricing tiers:
1. Small dataset (50–200 tracks, full stems): $10k–$40k upfront + 1–3% of product revenue tied to model outputs that materially derive from the dataset.
2. Mid dataset (200–1,000 tracks): $40k–$150k upfront + 1–5% rev share or per-call fee for commercial API usage.
3. Enterprise / exclusive dataset: $150k+ plus higher rev share (3–10%), or a multi-year licensing agreement with minimum guarantees.
• Why: Training data has higher multipliers because it powers models and creator tools. Buyers with strong funding (Higgsfield, Holywater) can afford larger upfronts but will push for broad rights — you must push back with limits and audit rights.

3) Exclusive sync or catalog sale

• Structure: Higher upfront, shorter term, stricter usage rights.
• Example pricing: 1–3 year exclusivity for a 250-track mini-catalog: $100k–$350k+ depending on demand and catalog uniqueness; include performance milestones and escalators.

4) Revenue-share + royalties hybrid

• Structure: Lower upfront ($5k–$25k), higher back-end (5–20% of net revenue from content using your tracks), with clear reporting.
• Why: Good for startups with limited cash but big growth curves — ask for minimum guarantees and timely reporting.

Key contract clauses every composer must insist on

AI-specific risks make certain clauses non-negotiable. These protect future earnings and your artistic control.

• Scope of use: Define permitted uses (training models, derivative generation, in-editor placement, broadcast) and expressly prohibit resale of raw stems unless separately licensed.
• Training limits: If allowing model training, restrict outputs that recreate or closely mimic your catalog; require a warranty that models won’t be used to generate identifiable copies without a separate license.
• Exclusivity & territory: Limit time (e.g., 12–36 months) and geography; prefer non-exclusive pilots.
• Attribution & provenance: Require discrete metadata/credits in UI where practical and technical provenance (hashing/watermarking) to track generated outputs back to your catalog.
• Reporting & audit rights: Monthly or quarterly usage reports and a right to audit, with agreed scope and frequency.
• Termination & takedown: Right to terminate for breach and require buyer to remove or disable derivative outputs within a set window.
• Payment terms & minimum guarantees: Net-30/45 for royalties, escrow for large upfronts, and minimum guarantees for multi-year deals.
• Indemnity & insurance: Narrow indemnities (buyer indemnifies for misuse) and require buyer to carry adequate tech/AI liability insurance.
• Attribution of moral rights: Clarify if works are to be altered and whether you retain moral attribution rights.

Negotiation playbook — step-by-step tactics

1. Start high but reasonable: Anchor with a number 20–40% above your target for room to negotiate. Use the pricing bands above.
2. Offer a pilot: Non-exclusive, 3-month, limited-track pilots close deals faster. Use performance-based escalators (if X plays in first 90 days, fee doubles or exclusive window offered).
3. Trade exclusivity for cash: If they want exclusivity, insist on higher upfronts + renewal bump clauses and short exclusivity windows.
4. Protect training rights: If they want dataset access, get higher upfronts, narrower training rights, and a clause preventing your music from being part of a public generative model unless separately compensated.
5. Ask for minimum guarantees: If they’re deep-pocketed (Higgsfield-style), secure minimum payouts or guaranteed placements across product lines.
6. Negotiate reporting cadence: Monthly is standard; insist on near-real-time reporting for in-app micro-usage if possible.
7. Escalators & audits: Add price escalators tied to scale and a right-to-audit clause with third-party auditors if necessary.
8. Walk-away triggers: Define behaviors that allow you to terminate (misuse, repeated reporting failures, distribution to third-party training marketplaces).

Red flags and deal breakers

• Blanket “perpetual, worldwide, irrevocable” training rights with no royalties or reporting.
• Refusal to provide usage reports or any audit mechanism.
• Demand to reassign publishing or writer shares as part of a catalog fee.
• Sublicensing to third-party datasets or marketplaces without your consent.
• No takedown procedure for outputs that infringe your moral rights or recreate your works.

Case study (composite, experience-based)

We advised a mid-career composer (composite of 8 real deals we’ve seen in 2024–2026) who split a 300-track catalog into pilot-friendly packages. They negotiated a 6-month, non-exclusive pilot with a well-funded vertical-video platform similar to Holywater: $30k upfront for 50 tracks + a 3% revenue share on creator monetization. After positive usage metrics the platform offered a 2‑year limited exclusivity for $200k plus higher rev share and stricter reporting. Key wins: upfront cash that funded the composer’s live streaming rig, audit rights that proved high usage (triggering their bonus), and a clause forcing attribution on creator pages. The composer kept the right to license non-overlapping tracks elsewhere and later leveraged the catalog into a placement in a branded miniseries for additional fees.

How to price by feature set (quick reference)

• Single hook (15–30s), non-exclusive: $75–$250
• Full track (30–180s), non-exclusive: $150–$500
• Stem pack (per track), non-exclusive: additional $100–$400 per track
• Dataset access (50–200 stems): $10k–$40k upfront
• Exclusive mini-catalog (100–300 tracks): $75k–$350k

2026 trends that change how you pitch

Three developments in late 2025–early 2026 matter for composers:

• Marketplace normalization: Human Native’s acquisition by Cloudflare signals enterprise moves to compensate creators — better leverage.
• Funding liquidity: Companies like Higgsfield and Holywater have fresh capital; they often prioritize rapid product growth and will pay for curated, ready-to-integrate catalogs.
• Regulatory & provenance momentum: Calls for provenance and metadata standards grew in 2025; buyers are increasingly expected to implement tracking and attribution — use this as leverage to demand traceable credits and reporting (multimodal media workflows).

Practical checklist before you hit send

1. Label all deliverables clearly (file names, version numbers).
2. Include a one-page licensing summary with your pitch.
3. Prepare a master metadata CSV (ISRC if available, track length, BPM, stems included).
4. Decide your walk-away price and non-negotiables (training limits, audit rights).
5. Have a contract template reviewed by entertainment counsel focusing on AI clauses.

Pitch template — one-pager (attach to outreach)

Include:

• Catalog summary: number of tracks, genres, average length
• Deliverables: WAVs, stems, dry mixes, source MIDI
• Suggested pilot: 20 tracks, 3 months, $X upfront, 2% rev share
• Licensing limits: non-exclusive, training allowed only for internal model improvements, not public distribution
• Reporting cadence and payment terms

Post-deal: How to monitor and scale the relationship

1. Set KPIs: number of placements, monthly plays, creator reuse rate.
2. Monitor usage via the agreed reports; trigger audits if numbers feel off.
3. Request quarterly review meetings to renegotiate escalators or add catalogs.
4. Use case studies (with permission) to market your catalog to other platforms.

Final negotiation tips from the frontline

• Put non-exclusivity first — it’s easy to scale from test to exclusivity, but hard to unwind a perpetual exclusive.
• Make the buyer commit to minimum guarantees when you give broad training rights.
• Insist on audit rights and clear KPIs; without them revenue shares are worthless.
• Use product integrations (editor placement, featured packs) as barter for better economics.
• When in doubt, ask for time-limited rights and renewal options tied to performance.

Actionable takeaways

• Segment your catalog into pilot-friendly packs (hooks, cinematic, dataset stems).
• Lead with a pilot: non-exclusive, 3 months, measurable KPIs.
• Price explicitly: use the bands above as anchors and trade exclusivity for cash.
• Protect training rights: specify how models can use your music and require attribution + reporting.

Closing — your next steps

If you’re ready: pick 10 tracks and build a 20-track pilot pack with stems and a one-page license summary. Use the email template above, ask for a 15-minute intro, and aim for a pilot with clear KPIs. In 2026, buyers like Higgsfield and Holywater will pay for ready-to-integrate music — but the difference between a fair deal and being undervalued is in how you package, price, and protect your work.

Want a custom outreach pack? Composer.live helps composers build pitch-ready pilot packs, metadata sheets and contract checklists tailored to AI buyers. Click through to create your first 20-track pilot and a negotiation-ready one-pager.

Related Reading

• Multimodal Media Workflows for Remote Creative Teams: Performance, Provenance, and Monetization (2026 Guide)
• AI Training Pipelines That Minimize Memory Footprint: Techniques & Tools
• Deepfake Risk Management: Policy and Consent Clauses for User-Generated Media
• Microdramas for Microlearning: Building Vertical Video Lessons Inspired by Holywater
• Brazil Auto Slump: What Q4 2025’s Downturn Signals for Global Auto Exports and Commodity Demand
• The Rise of Sensory Science: Could Receptor-Based Research Unlock Personalized Scalp Treatments?
• If Netflix Buys Warner Bros. Discovery: A Scenario Map for Shareholders and Competitors
• Create a Mini-Series Teaching Skincare Science — Episodes That Turn Complex Ingredients Into Stories
• Gifts for the Donut Lover: CES and Tech Finds That Actually Improve Your Baking Setup