Social scheduler built for real estate listings

The idea

Every listing an agent takes generates the same social lifecycle: just listed, feature highlights, open house, price change, just sold. Agents rebuild this by hand for every property on every platform, or pay an assistant to do it badly in Canva. Generic schedulers like Postiz prove people still pay for scheduling despite dozens of incumbents — but they hand agents a blank composer, which is exactly the work agents want to skip.

The wedge is modeling the listing as the object: paste a URL, get the photos and specs extracted, pick a campaign, and a week of posts appears on the calendar. A horizontal scheduler will never build listing scraping, lifecycle templates, or fair-housing copy checks — and that last one matters, because steering language in a caption is a real liability agents and brokerages actively worry about.

Who pays

Agents pay $29/mo and small brokerages $59/seat with approval workflows; agents already spend far more on lead-gen and photo services, and consistent social presence is how they justify their commission to sellers. Distribution: real estate agent Facebook groups and Instagram, brokerage office managers as multi-seat buyers, and content ranking for real estate social media templates.

MVP scope

• Listing scraper producing a confirmed structured record with photos
• Four campaign templates generating scheduled multi-platform posts via Claude
• Publishing to Instagram, Facebook, and LinkedIn with per-platform previews
• Fair-housing lint on all copy with explanations and rewrites
• Brand kit applied across campaigns
• Skip for v1: TikTok and YouTube, video generation, direct MLS/IDX integrations, DM management, analytics beyond basic post status

Build ListingLoop, a social scheduler where the unit of work is a listing, not a post. Stack: Next.js (App Router), Meta Graph API for Instagram and Facebook plus LinkedIn API, Claude API for copy generation and compliance checks, Postgres.

Core flow: agent pastes a listing URL (Zillow, Realtor.com, or brokerage site). A scraper pulls photos, price, beds/baths, square footage, and description into a structured listing record the agent confirms. Agent picks a campaign template — Just Listed (day 1 announcement, day 3 feature highlight, day 5 open house reminder), Open House, Price Improvement, Just Sold — and Claude generates platform-sized copy variants per post from the listing facts. The whole campaign lands on a calendar pre-scheduled; agent edits anything, hits approve, and posts publish automatically with the right photos per slot.

Pages: listings index, listing detail with campaign builder, calendar, post editor with per-platform preview, connected accounts, brand kit (headshot, brokerage logo, license number footer), billing.

Non-obvious details: run every generated caption through a fair-housing lint pass — a maintained blocklist plus Claude classification for steering language (family-friendly neighborhood, safe area, walking distance to churches) with flagged phrases explained and alternatives offered; when a listing status changes to sold on re-scrape, prompt the agent to fire the Just Sold campaign, which is the retention hook between listings.

Pricing: Agent $29/mo (1 user, 10 active listings), Team $59/mo per seat (shared brand kit, approval flow for unlicensed assistants).