Lumenary

What do they actually do

Lumenary makes Solar, a live web app you can sign up for today. Solar is a shared canvas where people and multiple AI agents build full‑stack software together. As you describe or sketch what you want, Solar proposes database schemas, backend Python logic, API endpoints, and UI pages, visualizes how they connect, and lets you deploy with one click. You can use Solar’s managed infrastructure for auth, storage, and databases, or connect your own systems (YC launch, YC company page, site).

The typical flow is: create a project, invite collaborators, sketch or describe features, review and remix the generated pieces on the canvas, test, and deploy. Solar combines LLM‑generated code with deterministic generation for core infrastructure so output is maintainable and suitable for production. Its agents can run unattended to refine work, read logs, and interact with the running app to resolve issues before deploy (YC launch, site).

Solar has opened public signups and publicized its launch through YC and press, indicating early availability and initial traction efforts (site, YC launch, press).

Who are their target customer(s)

• Solo founders and indie hackers building an MVP: They need to get a working, deployed product quickly but are slowed by stitching together design, code, and hosting across multiple tools (YC launch, site).
• Early‑stage startup product teams: They want to iterate features fast but lose time on handoffs between frontend, backend, and ops; collapsing design → code → infra → deploy into one shared workflow reduces this overhead (YC launch, YC page).
• Designers or PMs needing data‑driven prototypes that can become real apps: Clickable prototypes often get thrown away; they want outputs that are maintainable and deployable so work can be extended rather than rebuilt (YC launch, site).
• Small teams building internal tools without dedicated ops: They need reliable auth, storage, and data but get blocked by configuring infrastructure; deterministic generation for DBs/auth shortens setup and keeps apps maintainable (YC launch, YC page).
• Agencies and consultants delivering client MVPs: They must deliver working apps fast and hand off something clients can run and extend, not just demos (site, press).

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

• First 10: Founder‑led, white‑glove onboarding for solo founders and friendly teams (e.g., YC/early‑adopter networks), with hands‑on demos, setup help, and 2–3 detailed case studies to remove onboarding friction (YC launch, site).
• First 50: Run targeted build‑along workshops (Indie Hackers, HN, X), distribute ready‑to‑use templates, and follow up with personalized onboarding. In parallel, sign 3–5 small agencies to ship client MVPs and turn those into repeatable playbooks.
• First 100: Introduce a clear free tier or time‑limited credits, a small catalog of vertical templates, and product‑led onboarding (office hours, Discord, step‑by‑step guides). Scale partnerships with accelerators/agencies and add a referral credit program.

What is the rough total addressable market

Top-down context:

Solar sits at the intersection of AI‑assisted app development, low‑/no‑code app builders, and cloud IDEs—categories with large, established spend from startups, agencies, and internal tools teams.

Bottom-up calculation:

Initial wedge TAM: ~30,000 small teams and agencies globally adopting AI‑assisted app builders, averaging 4 seats each at ~$60/seat/month. That implies ~30,000 × 4 × $60 × 12 ≈ $86M in annual spend addressable by Solar.

Assumptions:

• Focus on early adopters among startups, internal tools teams, and agencies (~30k potential buyers).
• Average team size of 4 active builder seats per account.
• Seat‑based pricing around $60/month with standard SaaS retention.

Who are some of their notable competitors

• Anvil: Browser‑based full‑stack Python builder with drag‑and‑drop UI, built‑in data tables, and one‑click hosting; overlaps on fast shipping but centers on a Python IDE + visual editor rather than a multi‑agent canvas.
• Bubble: No‑code visual app platform for building web/native apps from a visual editor and AI prompts; strong for MVPs but trades off code export and traditional backend control versus generated, maintainable source.
• Retool: Low‑code platform for internal tools that connects to existing DBs/APIs and adds enterprise controls; optimized for internal workflows rather than collaborative, agent‑driven full‑stack product creation.
• Replit: Cloud IDE with real‑time collaboration, hosting, and AI agents that can generate/deploy projects; code‑first across many languages versus a visual canvas that maps relationships and deterministic infra.
• Appsmith: Open‑source low‑code platform for internal apps with self‑host and Git workflows; geared to internal tooling and widget+JS customization rather than multi‑agent, canvas‑driven product builds.