Backend tooling and automation pipelines.
Internal tools, custom integrations, scheduled jobs, monthly market-report generation, synthetic monitoring, MLS delta processing, document automation. The unglamorous infrastructure that turns a polished front-end into an actual operational system.
Front-end gets the glory. Backend silently breaks at 3am.
A custom IDX website looks great in screenshots. But behind the polish, real estate businesses depend on a dozen operational jobs that have to run reliably: MLS deltas every 15 minutes, monthly market report generation, lead enrichment from public records, automated comp pulls for new listings, weekly digest emails, transaction milestone reminders, document expiry alerts. When any of these fails silently, you find out from an angry client, not a monitoring system.
Most "real-estate platforms" don't do backend tooling at all. Sierra and BoomTown ship the website and call it a day. Luxury Presence ships the website and stops. If you want automation, you're hiring a fractional developer on Upwork, glueing things together with Zapier (and paying the per-task tax), or doing it manually in spreadsheets.
Custom backend tooling is just specialized software development — same skillset as the IDX site, same person, same support model. The unglamorous half of the stack that decides whether your business actually runs.
Scheduled jobs, custom dashboards, monitoring, integrations.
I've built and maintain a half-dozen production backend systems for SharonFronk.com: a 15-minute MLS delta sync, a daily MLS full-sweep, an hourly synthetic canary that fires fake events to test the system from outside, a bidirectional uptime watchdog where the website and CRM ping each other every 5 minutes, a monthly market-report email generator, a photo-takedown workflow for seller requests with full audit trail. None of these are exotic — they're just real-world operational software written carefully.
Backend work is scoped per-project. Typical engagements: $2K-$10K for a single tool or integration, with the option to bundle ongoing maintenance into a care subscription. For Reach Refine subscription clients (Refine and up), small backend work is included as part of the monthly edit allowance; bigger projects are scoped separately.
Backend tooling I've actually shipped.
- Scheduled jobs (cron)
- MLS deltas, full sweeps, market-report generation, lead enrichment, document expiry alerts. Runs on Vercel cron, GitHub Actions, or your infra. Failure alerts via Discord, Slack, or email.
- Synthetic canary monitoring
- Fake events fired hourly from known canary visitor IDs. CRM filters them from real surfaces but verifies they arrived. If zero canaries in 70 minutes, alert fires. The system tests itself.
- Bidirectional uptime watchdog
- Each app pings the other's /api/health every 5 minutes. 3 consecutive failures fires an alert. Recovery on first OK. Catches silent failures that don't throw HTTP errors.
- Custom internal admin tools
- Photo-takedown workflows, sharon's-picks curation, manual MLS data corrections, audit log browser, lead bulk-actions. UI built around your actual workflow, not a generic admin framework.
- Document automation
- Pull from DocuSign / Dotloop on signature. Trigger contractual milestones. Auto-populate disclosure packets. Expiry alerts. Document-of-record generation with proper PDF formatting.
- Custom integrations
- Whatever vendor doesn't have a Zapier connector. Phone systems (RingCentral, Dialpad), e-signature, transaction management, accounting (QuickBooks), brokerage commission engines. Direct API integration, no middleware tax.
Per-project, à la carte, or bundled into care.
Backend tooling is genuinely scoped per-project — a 30-minute scoping call gets you a fixed-price quote for the specific thing you need. Below are typical engagement sizes. Refine+ subscribers get small work (under 3 hours) included as part of monthly edits.
-
À la carte
$150-200/hr · single tool or integration, no minimum
- Scheduled job: $1K-$3K typical
- Custom dashboard: $2K-$5K typical
- Vendor API integration: $750-$2.5K
- Document automation: $2K-$8K
- No subscription required
-
Refine+ subscriber bundle
$1,495+/mo · small backend work included
- Work under 3 hours per month included in your edit allowance
- Bigger projects scoped at hourly rate (discount over à la carte)
- Same operator who runs your site
- Synthetic monitoring + uptime watchdog included on every Refine site by default
-
Care subscription
$799 – $1,995/mo · ongoing backend work for non-subscribers
- For clients on Reach (no subscription bundling) or buyout
- 8-20 hours of monthly backend work
- Includes monitoring + uptime + emergency response
- Cancel anytime after 3-month minimum
Backend tooling questions.
Can I just use Zapier for this?
Zapier is great for simple flows and rough integrations. It falls apart when (a) you need conditional logic that isn't trivial, (b) you need to integrate with a vendor that has no Zapier app, (c) the per-task pricing starts dominating at scale, or (d) you need observability into what's happening when things break. Custom backend tooling is the right call when Zapier becomes the bottleneck.
What's synthetic canary monitoring and why does it matter?
Most monitoring tools tell you when an HTTP request returns 500. They don't tell you when a request returns 200 but didn't actually do the thing it was supposed to do (silent failure). Synthetic canaries fire fake events from a known visitor ID at a regular interval; if the receiving system doesn't see them in the expected window, alert fires. It's the difference between 'the server is up' and 'the system actually works'.
Can backend tooling integrate with my existing IDX site (Sierra, kvCORE, etc.)?
Yes, via their API and webhooks. Sierra and kvCORE both expose decent APIs (kvCORE's is better-documented). The depth of integration depends on what their API allows — for some workflows we can do everything; for others we hit a ceiling and have to work around the platform's limits.
How does the cron infrastructure work?
Most projects run on Vercel cron (built into Vercel Pro, 60+ jobs allowed, simple JSON config). Heavier or more specialized workloads go on GitHub Actions or a dedicated cron service. Failure modes: every job logs to a centralized observability stack, failures trigger Discord/Slack alerts, retries are configurable per-job.
Can you maintain backend systems I had built by someone else?
Sometimes — depends heavily on code quality and tech stack. If it's a reasonable Node/Python/Go codebase with tests and documentation, yes. If it's undocumented PHP from 2014 held together with cron jobs and prayer, the honest call is usually to rebuild rather than maintain. Free 30-min audit will tell you which.
What about AI workflows — agents, RAG, etc.?
Yes — Anthropic Claude is in the standard toolkit. Real estate backend AI use cases I've built: lead intent scoring, automated property summary generation from MLS listings, weekly digest email writing, conversational document Q&A. Happy to scope a Claude-powered workflow during a discovery call.
Get a fixed-price quote in 48 hours.
Free 30-minute scoping call. Describe what you want built or what's currently broken; I'll give you a written fixed-price quote within 48 hours. Smaller projects often kick off the following week.
Book a backend scoping call