Open Source MLM Software

Criteria rather than names

We are not going to publish a list of recommended open source MLM projects, and the reason is practical rather than competitive.

Projects in this niche appear, get a burst of activity, and go quiet. A list we write today is misleading within a year, and if you adopted something partly because it appeared here, the consequence is yours and the cause is ours. Criteria do not expire. Apply the seven below to whatever you find on a code host today.

The seven checks

1. The licence. Read the actual file. This matters more than usual because you will modify the software and serve it over a network to your distributors:

AGPL is the one to read carefully. A distributor back office is software provided over a network, which is precisely the case AGPL was written to cover.

2. Maintenance signal. Commit recency, issue response time, whether reported security issues produced releases, and how many people have merge rights. A project with one maintainer is one person’s change of circumstances away from being your codebase.

3. The commission engine. The decisive check, and it takes ten minutes. Find the file that computes payout. Does it have tests? Do the tests contain worked examples with real numbers? Can a past period be recomputed and produce what it produced before?

If yes, someone thought hard about the part that matters. If no, every change you make to fit your plan is an unverified change to money-moving code.

4. Test coverage around payout specifically. Distinct from overall coverage. A project can be well tested everywhere except the calculation, which is the one place testing is not optional.

5. Security history. Has the project had vulnerabilities, and how were they handled? A project with disclosed and patched issues is in better shape than one with none reported, because the second usually means nobody has looked.

6. The data model, on two specific points.

• Does it store placement and sponsorship as separate structures? If your plan is width-capped and the project holds one tree, you cannot pay matching bonuses correctly, and adding the second tree later means reconstructing relationships you never recorded. The genealogy page explains why both are needed.
• Is volume materialised on write or computed on view? The second is fine at five hundred positions and unusable at fifty thousand, and changing it is a rewrite rather than a refactor.

7. Upgrade path. Once you have modified the commission logic to fit your plan, how do you take upstream updates? Usually the answer is that you stop taking them, at which point you own a fork and the security patching becomes yours.

What it actually costs

Free of licence cost. Not free of:

• hosting and infrastructure — modest,
• dependency patching, indefinitely,
• TLS certificate management,
• backups, and testing restores by performing them,
• monitoring and on-call,
• month-end capacity, on the day,
• and the engineering to implement whatever your plan needs that the project does not do.

That last item is usually the largest, and it is the same wall every free option runs into: the plan is specific to your company, and the project implements somebody else’s. The free MLM software page makes that argument in full.

What is almost always missing

Open source projects in this category are built around the interesting problem — the tree and the calculation. What they consistently lack is the record-keeping:

• agreement versioning, with the text a distributor accepted,
• income disclosure publishing,
• retail versus internal order classification at entry,
• an audit trail covering plan rule and rate changes,
• per-channel consent records.

None of these is technically difficult. All of them are impossible to backfill, because the data was never captured. A company that discovers this in year two cannot produce last year’s retail proportion at any price.

Where we land

Good for: learning the domain. Reading a working commission engine is genuinely instructive, and we would rather you understood plan mechanics than took our word for them.

Workable for: a very small single-market operation, a simple plan, and an engineer on the team who will still be there in eighteen months.

Poor for: paying thousands of independent contractors across two tax jurisdictions. Not because the code is bad — because the compliance layer is absent and unbackfillable.

If you are in the first two categories, our plan calculator and the mechanics pages on this site are free and will be useful to you regardless of what software you run.