Skip to content
Fair Software logoFair Software logo

Fair Score

What is Fair Score?

Fair Score is a 100-point system that Fair Software uses to assess how software products are built and run, considering fairness, transparency, and respect for users. We rate each app in our directory using 22 criteria in 6 sections. This score does not measure technical performance.

We assess each criterion through independent research in sources such as official documentation, privacy policies, security audit reports, public financial disclosures, and community activity. Every assessment comes with a written explanation and links to sources, so anyone can review and verify our evaluations.

Methodology

Every Fair Score evaluation follows a consistent research and documentation process:

  1. Primary source research.
    We begin with the product's own documentation: privacy policy, terms of service, security pages, pricing pages, and official blog or changelog. These are cross-referenced with third-party sources.
  2. Independent verification.
    Claims made by the company are verified against independent sources: audit reports, journalism, academic research, regulatory filings, and community reports. We do not take marketing copy at face value.
  3. Auditable justification.
    Each of the 22 assessments includes a written justification and direct links to the sources. Evaluations are stored as structured documents so that any reader can verify or dispute any individual assessment.
  4. Scoring discipline.
    Where evidence is ambiguous, we score conservatively — the benefit of the doubt goes to users, not the company. An app cannot earn points for a policy that contradicts its actual behavior.
  5. Regular review.
    Products change. Scores may be reviewed when significant changes occur (new ownership, major policy updates, security incidents, major new versions). The evaluation date is always shown on each product page.

The complete methodology — questions, scoring criteria, and rationale for every answer — is published openly at codeberg.org/gamito/fair-score under a CC BY-NC-SA 4.0 license.

Fair Software is not affiliated with, sponsored by, or compensated by any of the products it evaluates. Scores reflect our independent assessment based on publicly available information at the time of evaluation.

Rating Scale

Score80 – 100
What it means★★★★★ ExcellentExemplary across virtually all criteria. A reference product for ethical software.
Example
Signal 81
Score60 – 79
What it means★★★★ GoodStrong ethical profile with minor gaps. Trustworthy for most users.
Example
Vivaldi 71
Score40 – 59
What it means★★★ AcceptableMixed results. Some commendable practices alongside meaningful concerns.
Example
Ecosia 56
Score20 – 39
What it means★★ PoorSignificant ethical or privacy issues. Use with caution and awareness.
Example
Whatsapp 35
Score0 – 19
What it means★ HarmfulSerious concerns. The product works against users' interests in fundamental ways.
Example
ChatGPT 15

Questions

Fair Score is made up of 22 questions across 6 sections. Each question has a set of options with defined point values. The sections and their questions are listed below, with the rationale for each question.

Section 1 — Privacy & Data

Where user data lives, how it is protected, and whether it is used to train AI models.

1.1Where is user data stored?
The location of your data determines which country's laws govern access. In the US, laws such as the CLOUD Act may require companies to share information with authorities, regardless of your residence. Data stored in the EU or EEA is protected by the GDPR, which is among the world's strongest privacy laws. Storing data solely on your device eliminates these risks, as there is no server that can be subpoenaed, breached, or misused.
User's own device by default; data synchronization is optional and controlled by the user
6

The strongest possible position: nobody but you ever holds a copy of your data, so no jurisdiction, breach, or policy change can expose it.

EU/EEA, Switzerland, Iceland, Norway
5

These jurisdictions enforce GDPR or equivalent protections, including strong safeguards, independent oversight, and real penalties for misuse. It still involves trusting a third-party server, which is why it scores slightly below fully local storage.

Other democratic countries with strong data protection laws
4

Reasonable legal protections exist, but they may be weaker, less independently enforced, or more exposed to government access requests than GDPR-equivalent jurisdictions.

USA, China, countries with weak/uncertain data protection, or unknown location
0

Either the jurisdiction allows broad government access to user data (e.g., the US CLOUD Act, China's National Intelligence Law), or there is no clear information about where data is stored — an information gap that should concern any user.

1.2Is end-to-end encryption available for user content?
Encryption determines who can read your content. Without end-to-end encryption, the company and anyone who compromises its servers or receives a legal request can access your messages, files, or notes in plain text. End-to-end encryption means only you and your communication partner hold the keys, so even the provider cannot read your data.
Yes, full end-to-end encryption for user content by default or as an option
6

The provider has no technical ability to read your content, regardless of a breach, insider misuse, or a government request. This is the best protection available.

Encryption at rest and in transit only (no E2EE)
3

Your data is protected from outside eavesdroppers and physical theft. However, the company still holds the keys and can access your content if it chooses or is compelled to.

No meaningful encryption
0

Your content is stored and transmitted with little or no protection. This leaves it exposed to breaches, employees, and anyone who intercepts it.

1.3Does the company sell or share user data with third parties?
This is the difference between a company that respects your data and one that treats it as a product. Selling or sharing personal data fuels the surveillance advertising economy, profiling users without meaningful consent and often without their knowledge.
No, never (except for strictly necessary processors under data-processing agreements)
6

Your data stays where you put it. Necessary processors, like payment or hosting providers, are bound by contract and handle only what is needed to run the service.

Only anonymized or aggregated data for analytics or research
3

Individual users are not directly identifiable, but anonymization is often imperfect. Aggregated data can still reveal patterns about specific users when combined with other sources.

Yes, personal or behavioral data is shared or sold for advertising, profiling, or similar purposes
0

Your personal information becomes a commodity, traded to advertisers and data brokers, often used to build detailed profiles about you without meaningful consent.

1.4Does the product train AI models directly on user data (including prompts, files, metadata)?
Many products now use what you type, upload, or store to train their AI models without notice. This can expose sensitive or private information to a black-box model. Once your data is used for training, there is often no way to remove its influence.
No, never; or only with explicit, granular opt-in, off by default, and clearly explained
4

Your content is never used to train models unless you deliberately choose to allow it, with full understanding of what that means.

Yes, but there is a clear, easy opt-out that does not degrade the core service
2

The default favors the company, not the user. However, you have a real, unpenalized way to withdraw your data from training.

Yes, by default, with no meaningful opt-out (or only buried/obscure settings)
0

Your content trains the company's models whether you like it or not, and there is no genuine way to prevent it.

Section 2 — Openness & Transparency

License, auditability, open governance, and independent security assurance.

2.1What is the source code licence?
The licence determines who can inspect, modify, and reuse the code behind a product. Open licences let anyone verify what the software does and let the community fork the project if the vendor disappears or turns against users. Closed licences leave you dependent on the vendor's word.
OSI-approved open-source license for the core product
10

Anyone can inspect, audit, and fork the full product. No single vendor holds all the power over its future.

Open core (core open source, key features proprietary)
8

The base is transparent, but key functionality is locked behind a proprietary layer the community cannot inspect or fork.

Source-available (can be inspected, but the license is not truly open)
4

You can read the code, but licensing terms restrict modification and redistribution. Real transparency without the freedom to act on it.

Closed source (no public access to source)
0

There is no way to verify what the software actually does with your data or code. You are entirely dependent on the vendor's word.

2.2Has the software been independently security-audited or certified?
An independent audit catches vulnerabilities and bad practices before exploitation. A vendor's claims about its security are not verifiable alone; a public, third-party review lets you trust them.
Yes, a recent independent audit or relevant certification (within the last 3 years) with at least a public summary
5

An independent audit recently checked vulnerabilities; findings are public for anyone to review.

Yes, but the most recent audit or certification is older than 3 years
2

An audit exists, but software evolves quickly. An old review offers little assurance about the current codebase.

No known independent audits or certifications
0

No independent verification exists; you must simply trust the vendor's claims about security.

2.3For open-source projects, is the governance model documented and open to community participation?
Open governance means decisions about the project's direction are not made unilaterally behind closed doors. It gives the community a clear path to contribute, raise concerns, or fork the project if the vendor takes a bad direction.
Yes, governance is documented (e.g., maintainers and decision process), with contribution guidelines and a code of conduct
1

A clear path exists to contribute or influence decisions, reducing the risk that the project's direction serves only insiders.

No, or not applicable (closed-source or no documented governance)
0

No visible process exists; the project's direction is decided in private with no accountability to users or contributors.

Section 3 — Pricing Model

How users pay for the product, whether pricing relies on exploiting user data, and how fair and transparent prices are across regions.

3.1What is the primary pricing structure?
How you pay for a product shapes the incentives behind it. Subscriptions, especially "free" models that monetize your data, create ongoing pressure to extract more money or data. A one-time or lifetime price aligns the company's incentives with building something you own outright.
Free forever: no purchase, subscription, or data monetization of any kind
6

No payment, no hidden costs; your data is free with no strings attached.

Lifetime license / one-time purchase (no forced ongoing subscription for core use)
5

Pay once, own it forever with no recurring pressure and no risk of losing access after you stop paying.

Annual subscription (fair, predictable)
4

A recurring cost billed infrequently enough to remain predictable and easy to track.

Monthly subscription
2

A recurring commitment that's easy to lose track of and keeps constant pressure on the vendor to charge.

"Free" but monetizes user data (ads, profiling, data sales)
0

You pay with your data instead of money, often invisibly and without control over how it's used.

3.2Is there a genuinely functional free tier (without monetising user data)?
A free tier that works lets you use the core product without being forced to pay or give your data as the price of admission.
Yes, a fully functional free tier that does not monetize user data
4

Real needs can be met for free with no data monetization tradeoff.

Yes, but a limited free tier
2

A usable path exists, but restrictions may push you to pay or accept worse terms sooner than you'd like.

No meaningful free tier
0

There is no meaningful way to use it without paying, leaving out anyone unable or unwilling to.

3.3Is pricing transparent with no dark patterns?
Dark patterns manipulate users into paying more or agreeing to worse terms than they would choose with clear information. Transparent pricing respects your ability to make an informed decision.
Fully transparent pricing, clear plans, no manipulation
4

You know exactly what you're paying for and why, with no manipulation or hidden traps.

Minor issues or slightly confusing elements, but no clear manipulation
2

Pricing is generally understandable, but some ambiguity or friction may cause accidental overpayment.

Dark patterns present (e.g., tricking users into higher tiers, hidden charges)
0

Costs are deliberately obscured. Canceling or downgrading is made difficult, pressuring you to spend more than intended.

3.4Does the product use fair regional pricing that reflects local purchasing power without exploiting any region?
Flat global pricing set to US or EU income levels can make a product unaffordable in many parts of the world, excluding entire populations from access.
Clear, documented regional pricing based on local purchasing power
3

Prices adjust to local economic conditions, making the product broadly accessible wherever you live.

Some regional adjustments, but not clearly documented
2

Some accommodation exists, but lack of transparency makes it hard to know if you are getting a fair price.

Flat or opaque pricing with no fair regional adjustments
0

The same price applies everywhere regardless of local purchasing power. This prices out users in lower-income regions.

Section 4 — Ownership & Portability

User control over data and vendor lock-in, including export, self-hosting, offline use, cross-platform availability, and accessibility.

4.1Can users export ALL their data in an open format?
Without a full open-format export, your data is hostage to the vendor. Switching to another product means starting from zero or relying on a fragile unofficial workaround.
Yes, full export in open formats (JSON/CSV/XML, standard open formats)
6

You can leave at any time with everything intact, in formats any other tool can read.

Yes, but only in a proprietary format (requires the vendor's tools)
4

You can extract your data, but it is locked in a format only this vendor's tools understand. This complicates any real migration.

Partial export only (not all data, or heavily limited)
2

Some of your data can leave with you, but you will lose data or structure in the process.

No meaningful export
0

There is no real way to take your data elsewhere. You're stuck.

4.2Is self-hosting available?
Self-hosting, or running fully locally, means you aren't dependent on the vendor's servers staying online, priced fairly, or respecting your data — you control the infrastructure your data lives on.
Fully local: runs entirely on the user's device, no server or hosting required
6

Entirely under your control, with no dependency whatsoever on any company's infrastructure.

Yes, a full self-hosting option for the entire product (or a feature-complete edition)
5

You can run the whole thing on your own infrastructure if the vendor disappears or changes its terms.

Partial self-hosting (some features require the vendor's cloud)
4

Some independence exists, but core parts of the product still depend on the vendor's servers.

No self-hosting option
0

Total dependence on the vendor's infrastructure: if they shut down or change terms, you have no alternative.

4.3Does the software work offline?
Offline capability determines whether a network outage, an account suspension, or a vendor going dark can lock you out of your own work.
Yes, fully offline for core features
3

Full core functionality is available with zero dependency on a connection or the vendor's servers.

Yes, with sync (works offline and syncs when online)
1

Works offline but is designed for eventual reconnection and syncing. Pure offline resilience is more limited.

Requires an internet connection for core functionality
0

No connection means no access to your own data or work.

4.4Is the app available on multiple operating systems, including at least one non-proprietary/open platform?
Relying on a single closed platform ties your access to decisions by that platform's owner — app store bans, policy changes, deprecated OS versions. Availability on an open platform preserves an escape hatch beyond any one company's control.
Available on multiple platforms, including at least one open or non-proprietary platform (e.g., Linux, *BSD, F-Droid)
3

The widest reach with a genuine open-platform option not subject to a single company's app store rules.

Available on more than one platform, but only within closed ecosystems or with significantly reduced functionality outside the main platform
1

Some choice exists, but you remain fully dependent on proprietary app stores and their rules.

Available on a single proprietary platform only, with no public plans or APIs enabling broader access
0

Fully locked to one vendor's ecosystem, with no alternative if access is revoked or the platform changes its terms.

4.5Does the product demonstrate basic accessibility support (screen readers, keyboard navigation, contrast, captions, etc.)?
Accessibility determines whether people with disabilities can use the product at all. Screen reader support, keyboard navigation, and sufficient contrast are not optional extras — they are basic respect for all users.
Clear, documented accessibility support (policy or page) and evidence of basic WCAG-style features
2

A deliberate, verifiable effort to make the product usable for people with disabilities.

Some accessibility features are present but undocumented or clearly incomplete
1

Partial support exists, but you can't rely on it because it isn't documented or complete.

No visible accessibility features and no public commitment
0

No evidence the product considers users with disabilities at all.

Section 5 — Independence & Ethics

Ownership, independence from Big Tech, stability over time, and whether the company shows responsible behavior in ethics, advertising, and environmental impact.

5.1What is the company / project structure?
Corporate structure shapes incentives. Big Tech ownership and VC funding create pressure toward aggressive growth and monetization that can conflict with user interests; independent or non-profit structures are freer to prioritize users.
Non-profit, cooperative, or bootstrapped indie project (no external investors)
6

No outside investors demanding aggressive growth or an exit — incentives can stay aligned with users.

Small private company (no Big Tech parent, no large VC control)
4

Independent of the largest concentrations of platform power, though still a for-profit business.

VC-funded company
2

Investors expect large returns, which can eventually pressure the company toward growth or monetization tactics that harm users.

Owned or controlled by Big Tech (a large tech conglomerate with a surveillance-capitalism model)
0

Subject to the priorities of one of the largest, most powerful companies in tech, which rarely align with individual user interests.

5.2Has ownership changed significantly in the last 3 years?
Acquisitions often bring changed terms, new data-sharing practices, or a shift in priorities the original product did not anticipate. A track record of stable ownership signals trustworthiness.
No ownership change
4

The same ownership and incentives that earned your trust in the first place.

Minor change (e.g. small investor, internal restructuring)
3

Some change occurred, but the fundamental structure and incentives remain intact.

Acquired by private company (non-Big-Tech)
1

A new owner with different, unproven priorities and no track record yet.

Acquired by Big Tech
0

Now subject to the incentives and priorities of a Big Tech parent, regardless of what the original product promised.

5.3Does the company publish clear, public policies on responsible use, advertising, and AI?
Public commitments on how the product will and won't be used — advertising limits, no sale of data for surveillance, restrictions on harmful use — show whether a company has thought beyond legal minimums about its responsibility to users and society.
Clear, public ethical policies covering harmful uses, ads, and data/AI
3

Explicit, checkable commitments about how the company will and won't behave.

Some relevant ethical policies, but partial or limited
1

Some effort exists, but real gaps remain uncovered.

No clear public ethical policies beyond the legal minimum, or existing policies are contradicted by the company's actual practices
0

No commitments beyond what the law requires; nothing stops the company from acting against user interests short of illegality.

5.4Does the company provide public information or commitments regarding the environmental impact of its infrastructure?
Software running in data centers has a real energy and environmental footprint. A vendor's transparency and commitment — or a total lack of infrastructure to begin with — tells you whether that cost is being taken seriously or ignored.
No vendor-operated infrastructure at all: zero environmental impact from operations
2

Zero infrastructure means zero environmental footprint to account for — as good as any commitment could be.

Clear public information or commitments (e.g., green hosting providers, use of renewable energy, sustainability report)
2

The company has measured and disclosed its footprint, and made real commitments to reduce it.

Some indications (blog posts, partial mentions), but no coherent or up-to-date policy
1

Some effort is visible, but without a real policy to back it up, it's hard to know whether it's substantive or superficial.

No public information or commitments regarding environmental impact
0

No transparency whatsoever about the environmental costs of running the product.

Section 6 — Longevity & Community

How long the project has been actively maintained and how healthy and engaged its user and contributor community is.

6.1How long has the project been actively maintained?
Longevity is evidence that a product is stable and viable rather than a temporary experiment that might vanish along with your data or investment.
5+ years of continuous maintenance and releases
5

A long track record of active maintenance and staying power.

2–5 years of maintenance
3

Established enough to have shown some staying power, but not yet fully proven long-term.

Less than 2 years
1

Too new to have a track record — longevity and stability remain unproven.

6.2How active and engaged is the project's community and contributor base?
An active community and visible development show the project is alive, responsive to bugs and security issues, and unlikely to be abandoned suddenly.
Active community (forums/chats/issues), regular releases, visible roadmap
5

Signs of a healthy, ongoing project likely to keep improving and stay maintained.

Maintained but quiet (few public contributors, infrequent releases)
3

Still receiving updates but with limited visibility into the direction or health of the community.

Low activity, long gaps between releases, "abandoned" feel
0

Little sign of ongoing maintenance — a real risk that support and updates disappear entirely.