FablePool

A marketplace where customers pool capital to commission Claude Fable to build custom software in public.

Collaborative software development

What is FablePool?

FablePool inverts how custom software gets built. Instead of a single customer paying $50k for a consultant to deliver behind closed doors, FablePool lets five customers each contribute $10k to commission Claude Fable to design and build the same software in public view.

The entire development happens transparently. Customers watch Fable architect the system, write code, run tests, and ship. At the end, the pool negotiates licensing (closed-source, open-source, or tiered access). Every member gets the final software and complete visibility into how it was made.

Why this works:

Why This is Interesting

FablePool sits at the intersection of three market shifts:

1. The productization of AI agent work

AI agents are moving from research to revenue. But most businesses don't know how to turn agent labor into cash. FablePool shows one model: let agents build things, customers pool capital to fund it, the software becomes a shareable asset. This is productized AI labor.

2. Trust through transparency

Custom software is fundamentally a trust problem: Will the vendor deliver on time? Will it work? Will I regret this in six months? Public development solves this by making progress irrefutable. You see commits, test results, live architecture decisions. Estimation risk disappears.

3. The failure of traditional monetization

SaaS has commoditized. Services don't scale. Custom software is high-margin but doesn't scale either. FablePool bridges this paradox: it's a service (pool curation and coordination), but it delivers a software product (a scalable, shareable artifact).

Why a Landing Page Would Fail

FablePool cannot be communicated in a standard MVP landing page because the business model is process-heavy, not product-heavy. A landing page would require either:

FablePool needs an investment memo: honest, detailed, written for someone who can think in first principles rather than marketing narratives. This essay serves that purpose.

Market Landscape

Global custom software market: approximately $200 billion annually. FablePool's addressable segment: software projects under $100k where transparency is a competitive advantage and pooling reduces cost.

Initial beachheads:

Realistic 12-Month Product Evolution

Months 1-3: Manual Curation Phase

Wes manually forms 3-5 pools. A founder emails with a software need. Wes networks to find 2-3 other customers with the same problem. Wes negotiates licensing, timeline, and scope. Wes hands the spec to Fable. They build. Pool gets the software.

Goal: Prove (a) customers will pay, (b) Fable delivers quality, (c) pools are satisfied enough to tell peers.

Months 4-6: Basic Marketplace

Simple web application: customers post software needs and interest level. When similar needs cluster (three or more customers), Wes is alerted. He matches the pool, negotiates terms, launches the build.

Goal: Reduce Wes's manual networking overhead. Test whether customers self-organize into pools organically.

Months 7-12: Semi-Autonomous Matching

Smart pool formation: system automatically detects similar needs, suggests ideal team size, recommends licensing model (OSS vs. closed based on customer type), and generates standard contract terms. Wes reviews and approves 80% of pools without negotiation.

Goal: Run 2-3 parallel builds with Wes spending less than 10 hours per week on coordination.

Financial Model at Scale

FablePool takes 20% of total pool capital as a platform fee. A $50k build: platform earns $10k. Wes spends 8-12 hours on coordination and hand-off. Margin: $1-1.5k per hour. At three concurrent builds monthly, revenue annualizes to $360k.

Five Critical Questions

1. Will Wes enjoy pool curation?

FablePool's first 18 months depend entirely on manual pool formation. This is project management, not sales, not engineering. It requires negotiation skills, pattern recognition, and patience. If Wes prefers pure automation or wants to delegate early, FablePool stalls.

2. Can Fable actually deliver on FablePool's timeline?

The math assumes Fable executes a $50k project (equivalent to 6-8 weeks of senior engineer time) in 10-15 calendar days. Fable is fast, but has Wes validated this on complex, real-world software? Build complexity matters: a CRUD app takes 3 days; a distributed system takes 30.

3. What's the IP story?

If three customers fund a build, who owns the source code? If Customer A wants closed-source and Customer B wants open-source, the pool breaks. Is there a licensing model that keeps pools aligned? Restricted OSS? Shared source? Revenue-share? This is the unsolved problem.

4. Is Wes's network large enough?

FablePool assumes Wes knows 20+ potential customers with $10-50k software needs. Does he? Or is customer acquisition a cold-start problem that kills the hand-curated model?

5. What's the failure playbook?

Fable misses a deadline. The software ships with critical bugs. A customer wants to exit mid-build. What's the SLA? The refund policy? The remediation process? Lack of clarity here destroys trust after the first incident.

Before Building an MVP

FablePool is not ready for a marketplace web app. Premature infrastructure is expensive and can collapse under bad assumptions. Instead:

  1. Hand-form two manual pools end-to-end. Complete one build from spec to delivery. Did Fable deliver? Did the pool stay cohesive? Would members recommend it?
  2. Document metrics. Timeline, scope changes, bug fixes, customer satisfaction score. Are pools repeatable?
  3. Test organic demand. Can you recruit the third and fourth pool without cold outreach, or do prospects require active sales?
  4. Only then design the marketplace. If manual pools are working, scale with automation.

If any of these fail, FablePool was saved from a six-month engineering bet. If all three succeed, the marketplace is built from proven demand.

Why This Matters

FablePool tests whether AI agents can be vertically integrated into business processes. It answers: Can humans and agents collaborate to deliver software that humans pay for? Can transparency replace contracts as a trust mechanism? Can pooling solve the software cost problem?

If it works, the model extends to other domains: pooled data analysis, pooled design, pooled infrastructure. If it fails, Wes learns exactly which assumptions were wrong.