Developer tools decision

NEURAL DRAFT

Self-hosting is realistic: the project publishes an MIT MCP server repo and the product is largely an API+integrations stack, so a capable technical user can stand up and run a useful replacement and only give up hosted billing, managed scaling, and vendor support.

Visit website

Built by Vedran Balagovic 💙, who ships 5 products in this index

You pay

$25/mo

$300/yr

Read off the official pricing page.

You’d pay instead

$20one-off6 h to build

$100/mo6 h/mo upkeep

On cash alone, building overtakes the subscription at 5 seats.

What a replacement has to do

  • Generate a blog post via an LLM, store content in a CMS, publish to a public site, create scheduled social posts, accept bookings, and create a Stripe checkout session.

What it still won’t have

  • Hosted credit billing, top-ups, and hard-stop credit enforcement
  • Turnkey managed integrations and uptime SLAs for social connectors and Stripe Connect
  • Built-in paid image/video generation pipelines tied to vendor credits (e.g., Gemini Nano Banana)
  • Hosted dashboard, managed scaling, and vendor support/priority support

What remains hard

  • Product polish and ongoing maintenance
Read the build prompt

First-year cost

Keep paying

Paying is—cheaper in year one.

On cash alone, building overtakes the subscription at 5 seats.

Paid seatsseats

Money you would actually spend

Keep paying
—

Subscription price × seats × 12

Build it
—

AI build —APIs + hosting —

Time you would spend

—

—

What you would spend

What we assumed

The verdict above measures whether you could build it. This one is only about money.

Runnable build prompt

Not run yet
Build a self-hostable Neural Draft backend using Node.js (NestJS or Express), a Postgres database, a Redis-backed background worker (BullMQ), and a simple React admin UI. In scope: API endpoints for content/blog/social/bookings/commerce; database schema and migrations; LLM integration adapter (OpenAI/Anthropic)—abstract as a pluggable provider; social API adapter for scheduling (Twitter/X, LinkedIn) with queueing; Stripe integration with checkout session and optional Connect flow; MCP-compatible command-line server component to provide MCP responses to Claude/Cursor; background job processing for generation, translation, image requests, and scheduled posts; basic auth and API-key based access; CI tests for API endpoints and worker jobs; environment-based configuration, Dockerfiles, and a deployment guide for a single VPS (DigitalOcean) plus managed Postgres. Out of scope: multi-tenant SaaS billing UI, hosted credit marketplace, proprietary image model hosting, enterprise white-label SLAs. Provide error handling, retries for transient failures, logs, health checks, and unit/integration tests covering core flows.
How we checked4 sources · 1/1 runs agreed · evidence score 99

How the score was reached

  • Self-host verdict base92
  • An open-source build was found+5
  • 4 cited sources+3
  • Price verified on pricing page+3
  • 1/1 assessment runs agreed+4
  • Evidence score99

The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.

How scoring works →

Cited sources · 4

Every page the run actually retrieved.

Integrity checks

What held up, and what did not.

✓ Price read off the page✓ 1 independent runs, one answer✓ Citations limited to fetched pages! 1 moat recorded