# Flare ships five targets from one Kotlin Multiplatform tree

> Flare is a social client that merges Mastodon, Misskey, Bluesky, X, Pixiv, Fanbox, Nostr and RSS into one timeline, written in Kotlin Multiplatform with native apps for five desktop and mobile platforms plus a web build. The roadmap in the README is more revealing than the feature list.

**DimensionDev/Flare** — Browse Mastodon, Bluesky, X, Misskey, Nostr, Pixiv, Fanbox and RSS all in one app. One timeline, all your accounts, cross-post everywhere.

- Repository: https://github.com/DimensionDev/Flare
- Website: https://flareapp.moe/
- Stars: 1,642 · Forks: 77
- Language: Kotlin
- License: AGPL-3.0
- Published: 2026-10-16 · Updated: 2026-10-16 · Language: en
- Canonical page: https://hysenlabs.com/projects/dimensiondev-flare

## Eight networks behind one timeline

The premise is aggregation without a server. Flare brings Mastodon, Misskey, Bluesky, X, Pixiv, Fanbox, Nostr and RSS into a unified timeline, and each of those has its own model: a federated microblog, an ActivityPub implementation with its own extensions, a decentralized protocol with a very different data model, a closed platform, two Japanese creator platforms, a decentralized relay protocol and plain feeds.

The mixed timeline is the feature that decides whether this is useful. The README describes it as merging content from multiple accounts and platforms into a coherent feed, reducing context switching between apps. And the roadmap shows grouped mixed timeline already checked off, so the grouping work is done rather than promised.

The feature list goes beyond a reader. Polls, lists, bookmarks and favorites are supported, along with Misskey antennas, Bluesky feeds and direct messages, and RSS management. Cross-posting is the counterpart for people who publish rather than read: users can publish to multiple platforms at once, which the README frames as making it practical for creators and heavy social media users.

Direct messages are supported, which is worth noting for a client described as privacy-first: the privacy claims are about data handling by the client, not about the security of the underlying networks.

## The repository is one Gradle build with a target per platform

Flare is built with Kotlin Multiplatform and shares its core logic across Android, iOS, macOS, Windows and Linux, with a web client as well. The platform badges state the minimum versions directly: Android 8.0, iOS 17.0, Windows 10 build 1809, macOS Sonoma, and Linux as an AppImage.

The top level shows how that maps onto code. app/ is the Android application, iosApp/ and appleApp/ plus apple-shared/ cover the Apple side, desktopApp/ covers the desktop targets, and web/, web-shared/ and web-presenter-processor/ cover the web build. The shared core lives in shared/ and social/, with compose-ui/ for the interface layer and feature/ for feature modules.

Two entries in that list are unusual and worth understanding. build-logic/ is the convention plugin layer for the Gradle build, which is how one project ships five targets without five copies of configuration. And web-presenter-processor/ is a build-time code processor, which suggests the web target reuses presentation logic through generation rather than through a hand-written binding.

Publishing is standard multi-channel: the Microsoft Store, the App Store, Google Play, F-Droid and an AppImage from the releases page, with fdroid.properties and metadata/ at the root for the F-Droid packaging.

## The roadmap is a list of unfinished work

The roadmap section separates shipped items from planned ones with checkboxes, and the split is more informative than the feature list.

Done: grouped mixed timeline, crossposting for repost, support for Nostr, the desktop client and the web client.

Not done: showing an instance's announcement, auto thread, personal trends of the day, quick reply, support for Meta Threads and support for a Discourse forum.

The AI features appear in two places, which is a small inconsistency worth noting. The feature list describes a built-in agent already available, with summaries, post and profile insights, chat, and queries against local cache and history, configurable with an OpenAI-compatible provider or an on-device model. The roadmap still lists AI powered features with quick reply and personal trends unchecked. Both can be true, since the agent exists and specific features are pending, but a reader comparing the two sections will pause on it.

Auto thread and Meta Threads support are the two gaps most likely to be noticed by a daily user. The Discourse forum item tells you the long tail of forums is not covered yet, which is the boundary of the aggregation story: it reaches networks with structured APIs, not arbitrary web pages.

## Privacy-first means local, and it does not mean private

The privacy claim is stated as a feature: as a FOSS client, Flare emphasises user control with features like anonymous mode, local filtering, local history and transparent data handling.

Read carefully, that is a claim about where work happens rather than about cryptography. Local filtering and local history mean your rules and your read state live on your device rather than on a server, which is a real advantage over the algorithmic feeds it replaces. Anonymous mode is a presentation and network choice.

What it does not cover is the networks themselves. These features apply to the client and do not establish privacy guarantees for the networks behind it. Flare being AGPL-3.0 and FOSS says you can audit the client, not that the platforms behind it are private.

The AI agent deserves the same scrutiny. Flare supports both an OpenAI-compatible provider and an on-device model.

## A community project with channels, not support

Support and contribution follow the pattern of a volunteer project. There is a Telegram group, a Discord server, a Crowdin project for translations with crowdin.yml at the repository root, and a DeepWiki badge. Contributing is handled through CONTRIBUTING.md, with a separate CODE_OF_CONDUCT.md.

The licence is AGPL-3.0, which for a client application means that if you modify it and host the modified version as a network service, you owe source to your users. For most individuals running the app that constraint never comes up. For an organisation that wants to fork Flare and put an internal instance behind a login, it is the thing to think about first.

Release cadence is regular: 1.7.7 on 2026-09-30, 1.7.6 on 2026-09-18 and 1.7.5 on 2026-09-04, with the last push on 2026-10-15. Patch numbers on a 1.x line with several releases a month is a project being shipped to end users continuously rather than maintained occasionally, which is what you want from a client whose value depends on tracking network changes you did not control.

## Conclusion

Flare suits a reader or creator who lives in several of these networks and wants one timeline plus one place to post, and the AGPL-3.0 licence makes it straightforward to run your own build if you need to change something. It is a poor fit for a team that wants a maintained platform client, since the project is community run with a public roadmap and no support commitment. Before adopting, read the roadmap items still unchecked, especially Support for Discourse forum and Auto thread, and decide whether the AI agent's OpenAI-compatible or on-device configuration fits your privacy expectations.

## FAQ

### Which networks does Flare support?

Flare brings Mastodon, Misskey, Bluesky, X, Pixiv, Fanbox, Nostr and RSS into one timeline, and also supports polls, lists, bookmarks and favorites, Misskey antennas, Bluesky feeds and direct messages.

### Which platforms does Flare run on?

It is built with Kotlin Multiplatform and ships for Android 8.0 or newer, iOS 17.0 or newer, Windows 10 build 1809, macOS Sonoma as a desktop app, Linux as an AppImage, and the web.

### What AI features does Flare have?

The README describes a built-in agent that can be configured with an OpenAI-compatible provider or an on-device model, providing summaries, post and profile insights, chat, and queries over local cache and history. Individual AI features such as quick reply remain unchecked on the roadmap.

### What licence is Flare released under?

Flare is licensed under AGPL-3.0, with the licence file at the repository root. Builds are also distributed through Google Play, F-Droid, the App Store, the Microsoft Store and GitHub releases as an AppImage.

## Sources

- [DimensionDev/Flare on GitHub](https://github.com/DimensionDev/Flare)
- [License: AGPL-3.0](https://github.com/DimensionDev/Flare/blob/master/LICENSE)
- [Project website](https://flareapp.moe/)
- [README](https://github.com/DimensionDev/Flare/blob/master/README.md)
- [Releases](https://github.com/DimensionDev/Flare/releases)

---

Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/dimensiondev-flare