The social media landscape is marked by a high degree of centralized power, with significant consequences for free and democratic societies. A small number of corporations now mediate how billions of people communicate, discover news, share images, host media, and collaborate on software. The harms are not accidental side effects: they follow directly from business models built on surveillance, engagement maximization, and closed ecosystems.

State-level censorship is not a viable solution, yet alternatives exist that allow different communities to operate their own platforms, whether peer-to-peer or federated. These models enable groups to apply their own social rules without relying on a single point of control, and always with the option to migrate to another instance.

This dossier presents NGI-funded alternatives to mainstream social media. Each project below addresses a specific class of risk posed by centralized platforms. Together, they show that the benefits of social connection can be preserved without accepting surveillance, addiction, exploitative design, or platform lock-in as the price of participation.

Understanding what is at stake is the first step toward meaningful alternatives. The dominant platforms expose users, communities, and societies to interconnected risks:

Surveillance and profiling. Platforms collect behavioural data far beyond what users consciously share: who you follow, when you are online, what you pause on, and who is in your network. This data powers advertising profiles, can be subpoenaed or leaked, and creates a permanent record that users neither own nor control.

Addictive design and mental health. Infinite scroll, variable-reward notifications, and algorithmic feeds are engineered to maximize time on platform. Research links heavy use to anxiety, depression, sleep disruption, and reduced attention span, with adolescents especially vulnerable.

Harassment and unsafe exposure. Global reach combined with weak or opaque moderation leaves victims with little recourse. Reporting systems are slow, inconsistent, and designed to protect the platform's reputation rather than the people harmed.

Opaque feeds and engagement optimization. Proprietary ranking systems decide what billions of people see, without transparency or user control. Feeds are tuned to maximize time on platform rather than user wellbeing. Users cannot inspect, adjust, or replace the logic that orders their timeline, and the incentives behind it are commercial, not community-driven.

Platform lock-in and loss of digital sovereignty. Years of posts, contacts, and community history are trapped inside proprietary silos. Leaving a platform means starting over. Creators depend on algorithms they cannot inspect for income and reach.

Monopolization of public and cultural infrastructure. A handful of companies control not only social networking but also image sharing, podcast distribution, link aggregation, and developer collaboration. Outages, policy changes, or account bans can silence individuals and communities overnight.

Environmental cost. Data centres powering social media and AI-generated content consume enormous energy. Engagement-maximizing design further increases compute demand through constant ranking, recommendation, and content generation.

Interoperability failure. Walled gardens prevent users on one service from reaching friends on another without surrendering more data to a new monopoly. Federation promises openness, but only if implementations actually interoperate.

These risks are no longer theoretical. Social media affects adolescent wellbeing, personal safety, and the livelihoods of journalists, activists, and creators. The longer communities depend on centralized infrastructure, the harder migration becomes and the more power accumulates in hands that are accountable only to shareholders.

For individuals, the cost is personal safety, privacy, and mental health. For communities, the cost is dependence on infrastructure they do not control and cannot leave without losing years of connections and content. For the planet, the cost is rising energy use driven by engagement-maximizing design and large-scale content processing.

Waiting for regulation alone is insufficient. Laws lag behind technology, vary by jurisdiction, and cannot easily undo network effects once a platform becomes the default place where a community lives. Practical alternatives that communities can run themselves are needed now, not after the next scandal or outage.

The projects in this dossier each target specific harms of centralized social media. None is a drop-in replacement for every proprietary platform, but together they cover the full stack of social infrastructure:

	ActivityPods 3.0 confronts surveillance and lock-in by giving users encrypted, portable data stores (Solid-compatible Pods) combined with Fediverse communication. Users control where their data lives and which apps may access it, rather than trusting a single company's servers.
	Bana reduces exposure to mass-scale profiling and harassment by focusing on private, personal social networking. Federation means users are not trapped on one operator's servers and can move if governance changes.
	Bonfire directly rejects the engagement-driven model: no ads, no surveillance, no opaque recommendation algorithms. Communities define their own moderation rules, permission boundaries, and feed logic, and can migrate all their data to another instance.
	Castopod (including Castopod Mobile and Castopod Plugins) counters creator dependence on proprietary podcast platforms such as Spotify. Podcasters keep direct relationships with listeners through open protocols, can self-host, and extend their server with plugins for local legal and community needs instead of accepting one global policy.
	Fediverse Test Framework addresses the interoperability risk that could turn federation back into fragmentation. By testing how ActivityPub messages are parsed and rendered across implementations, it helps developers build software that actually connects rather than creating new silos.
	ForgeFlux tackles monopoly risk in developer social infrastructure. Software forges such as GitHub function as social platforms for millions of developers; federation via ActivityPub reduces dependence on a single corporate operator for collaboration and visibility.
	Lemmy offers a decentralized alternative to proprietary link aggregators like Reddit. Communities set their own moderation policies, votes are transparent, and ranking is not driven by a corporate engagement engine tied to surveillance and lock-in.
	PixelDroid provides a mobile path away from surveillance-based image sharing on platforms such as Instagram. Users can connect to federated Pixelfed instances, hold multiple accounts across servers, and share photos without feeding a centralized behavioural profile.
	Spritely addresses a risk that technical alternatives alone cannot solve: if security and privacy features are too hard to use, people default back to convenient centralized apps. It works on making decentralized social media safe and usable through better interface design.

Free/Libre and Open Source Software (FLOSS) alternatives are an essential part of a healthier digital future. They show that social platforms can exist without ads, addictive design, or surveillance business models, and that a different kind of digital world is already taking shape.

They also provide a real fallback when dominant systems fail or become harmful, much like skyscrapers rely on stairs even if we rarely use them. FLOSS tools allow communities to shape how their digital spaces work, encouraging experimentation and democratic decision-making.

Because they are not driven by profit, they can afford to support slower, more meaningful interactions and less computation-intensive practices. They can also be installed and operated locally, from a basement server to a neighborhood association, a municipal facility, or a national sustainable datacenter.

The projects above are not a complete escape from every risk overnight. Federated systems carry their own challenges: instance administrators can still make bad moderation decisions, smaller servers may lack resources, and users must learn new tools. But the fundamental asymmetry changes: power is distributed, code is inspectable, and exit remains possible. That is the minimum condition for social media that serves people rather than extracting from them.

Browse the NGI-funded projects in this dossier below.