Post-SaaS · the argument

Renting made sense when software was expensive. It isn't anymore.

SaaS existed because building and operating software cost real money. AI collapsed both costs. What comes next is ownership that doesn't rot: MIT code, your data, your infrastructure, and an upstream paid to keep refining it.


01

Why SaaS existed

Software as a service was never about the software. It was about who paid for the engineers.

Building a support desk or a status page took a team. Operating it took another. Nobody sensible wanted to fund either for a tool that was a means to an end, so the market did what it does: one company built the thing once and rented it to everyone. You gave up your data, your price, and your say over the roadmap. In exchange you never had to think about it.

That was a good trade for twenty years. It rested on one assumption: that writing and running software stays expensive. That assumption is gone. A capable agent now builds, tests, and maintains a complete product in the time a procurement call used to take. When the cost of making the thing approaches zero, the rent stops being a bargain and starts being a toll.

02

Ownership used to mean maintenance

The thing SaaS really sold you out of was the burden. Self-hosting meant a repo, a server, and a Saturday.

Open source offers editable starting points, often with strong maintenance and support. Capable agents can adapt and maintain them. Run Edit Run must earn its place by saving the total effort of reaching and keeping the owner’s desired product.

The owner’s agent can change any source and adapt useful upstream improvements, including after a deep rewrite. Requirements and verification help it preserve what matters. Optional management is planned to make that work convenient; it must demonstrate the resulting product and continued independence.

03

What ownership means here

Written baseline requirements guide the products. Editions record applicable scope and intentional differences; inspect release evidence for implemented behavior and test coverage.

The requirements below come from the shared BASELINE.md. An edition’s release evidence must show which requirements apply and what has been verified. These requirements do not themselves promise a management service or production readiness.

BASE-DATA-005
Independence.

The deployment makes no request to the marketplace or the seed author. There is no telemetry, phone-home, licence check, or kill switch.

Requirement · check release evidence
BASE-DATA-001
Export and import.

The required outcome is complete export and restoration of records and attachments. Use the selected edition’s actual commands and inspect its verified coverage.

Requirement · check release evidence
BASE-DATA-004
Outbound flows.

User data leaves the deployment only through externals declared in seed.json → externals. No other outbound request carries user data.

Requirement · check release evidence
BASE-DATA-002
Retention.

No user data is deleted automatically. Only sessions and rate-limit state expire.

Requirement · check release evidence
BASE-PUBLIC-001
Privacy.

Public surfaces (widgets, booking pages, forms, status pages) set no cookies, load nothing from third-party domains, and send no visitor data before the visitor's first interaction.

Requirement · check release evidence
BASE-SECRET-001
No leakage.

No secret appears in the client bundle, in logs, or in an error response.

Requirement · check release evidence
BASE-OPS-004
Failure is visible.

A job that fails after its retries is marked failed where an operator can see it. Nothing is dropped silently.

Requirement · check release evidence
BASE-OPS-005
Audit.

Every destructive action and every automated action that reaches a user records the actor, or the feature that acted, and the time. Audit records survive deletion of what they describe.

Requirement · check release evidence

These are product requirements for the selected edition. Temporary management previews are evaluation environments with separate expiry and cleanup arrangements; they are not production data storage. Authorised management and editor features also have their own disclosed data flows. See the service terms and privacy policy.

The full baseline runs to twenty-two clauses across access, input, secrets, data, public surfaces, and operations. Each seed's detail page renders its own product contract on top, clause by clause.

04

Why pay for MIT code

Because the code was never the expensive part. The refinement is.

An edition should provide a cohesive product: useful defaults, integrated source, recorded requirements and meaningful tests. General-purpose improvements we have rights to publish become shared source and verification cases, without publishing private customer code or data. HappyToHelp is the first launch path: standalone is free MIT, while Assisted and Managed charge for deployment, management and update work.

Published editions include their selected foundation source under MIT. Shared refinement and optional management must earn their value against other open-source products, hosted services and an agent building from scratch.

Subscription SaaS

Renting

  • Their data centre, their export policy
  • Their price, revised yearly
  • Their roadmap, their sunset date
  • Nobody can kill it — except them

Convenient until it isn't.

A seed

Owning

  • MIT code, your cloud, your data
  • Requirements and recorded verification evidence
  • Useful fixes adapted by your agent
  • Shared refinement; management planned

Yours, and still getting better.

Other open source

Adapting

  • Editable starting products and libraries
  • Tests and support vary by project
  • Upstream work your agent can adopt
  • Compare the full adaptation effort

A useful alternative to compare.

See what owning looks like.

Fourteen complete products, each with its contract on the table.