Mosaic

What do they actually do

Mosaic is a web-based video editor in public beta that lets you build reusable editing workflows on a visual canvas. Users drag Tiles (e.g., Rough Cut, B‑Roll, Captions, Music, Motion Graphics, Upscale) to assemble an Agent, then run it to automatically apply those steps and branch into multiple edited variants in parallel (product page, Launch HN).

After a run, you can review results, make tweaks via a chat copilot or a frame‑accurate timeline editor, and export XML for handoff to Premiere, Resolve, or Final Cut. The product also plugs into multiple generative and enhancement models for video, voice, music, and upscaling, which you can include as steps in an Agent (product page).

For teams, Mosaic provides an API, triggers, and webhooks so uploads can automatically kick off configured workflows; runs and versions are trackable, and node‑run billing is used today to cover model costs. The company notes it is not yet a fully autonomous, “set‑it‑and‑forget‑it” editor—users should expect to do final polish in the timeline or their NLE (API docs, Launch HN, product page).

Who are their target customer(s)

• Solo creators and independent YouTubers who publish regularly: They spend time on repetitive trimming, captions, B‑roll placement, and producing multiple platform‑specific cuts; they want to automate these repeatable steps (product page).
• Social media managers and small agencies repurposing long videos into many short variants: Creating captions, edits, and A/B variants for each channel is slow and error‑prone; they need automation and branching to produce consistent variants quickly (product page).
• In‑house marketing/video teams at startups and SMBs: Manual handoffs and repetitive tasks slow their pipeline; they need triggers, webhooks, and an API so new uploads automatically run a configured workflow (API docs).
• Freelance editors and small post houses: They repeat the same rough‑cut work and still finish in Premiere/Resolve/Final Cut; they want reliable automated rough cuts and clean XML exports to finish locally (product page).
• Course creators and educational teams: They spend hours turning lectures into clips, captions, and highlights across many lessons; they need repeatable steps to reduce manual chopping and captioning.

How would they acquire their first 10, 50, and 100 customers

• First 10: High‑touch onboarding of creators from the founders’ network, YC contacts, and Launch HN; offer free credits and 1:1 sessions to build/tweak a first Agent, gather feedback, and convert to paid via node‑run billing (Launch HN, YC listing).
• First 50: Seed communities with workshops/AMAs and publish short case studies showing time saved and XML handoff; pair with outreach to small agencies and freelancers for paid pilots, plus template Agents and trial credits to drive self‑serve signups (product page).
• First 100: Lean into integrations (YouTube/cloud ingest, webhooks, API runs) to demonstrate automated pipelines; run referrals and 30–60 day paid pilots with in‑house teams, converting pilots to monthly plans using templates and API/triggers as proof points (API docs, product page).

What is the rough total addressable market

Top-down context:

Global video‑editing software is on the order of a few billion dollars (around ~$2.4B in 2024), while broader creative/digital content creation tools are in the high single‑ to low‑tens of billions (TBRC, Grand View Research, Fortune Business Insights).

Bottom-up calculation:

Using ~207M global creators as a base, 0.5% converting at $50/yr implies ~$52M ARR; 1% converting at $120/yr implies ~$248M ARR (Forbes/Statista).

Assumptions:

• Global creator base ≈ 207M
• Paid conversion range: 0.5%–1% of creators
• ARPU range: $50–$120 per year

Who are some of their notable competitors

• Descript: Transcript‑first editor that automates cuts, captions, and audio cleanup; overlaps on automated rough cuts and captions but centers on text‑based editing and voice cloning rather than canvas‑built agents.
• Runway: AI‑first video suite focused on generative video and in‑browser VFX; overlaps on AI generation/enhancements but emphasizes creative generation over programmable, branching agent workflows.
• Kapwing: Browser editor with one‑click AI (auto‑subtitles, resize, text‑to‑video); strong for quick social assets and templates but less focused on agent pipelines or XML handoff to pro NLEs.
• VEED: Online editor bundling auto‑subtitles, AI clips, and brand kits; good for social repurposing at team scale, but mainly a conventional project/template editor rather than branching agents on a canvas.
• Pictory: Turns long videos or text into short clips and summaries automatically; optimized for long‑to‑short repurposing, not user‑defined, inspectable agent pipelines or API‑driven branching runs.