Overview
AFFiNE is a local-first workspace that does not treat Notion-style documents and Miro-style whiteboards as separate tools. The same block data can move between Page and Edgeless modes, mixing rich text, sticky notes, linked pages, databases, shapes, slides, and embedded content in one workspace. Data ownership starts on the user's device while real-time sync and collaboration can extend the workspace across web, desktop, and self-hosted infrastructure, making AFFiNE as much a knowledge-base substrate as a note-taking application.
Features and best fit
Based on official documentation; not hands-on tested · Content checked:
Move blocks between structured pages and a freeform canvas
AFFiNE's defining feature is a shared block model for document editing and whiteboarding. Content created in a page can be placed on the Edgeless canvas alongside rich text, sticky notes, linked pages, multi-view databases, web embeds, shapes, and slides. That makes workflows such as turning meeting notes into a relationship map, expanding a specification into a planning board, or visualizing database-backed information possible without treating the canvas as a disconnected export surface.
Sources: [1]
Keep data local-first with CRDT-oriented sync and collaboration
The project explicitly positions AFFiNE as local-first, emphasizing that users retain data on disk independently of the cloud. Its upstream stack includes BlockSuite, y-octo, OctoBase, and Yjs, providing the local-first and CRDT-oriented foundations behind synchronization and collaboration across web and cross-platform clients. This makes it relevant when offline ownership is important but teams still want to add real-time collaboration when needed.
Sources: [1]
Use AI for summaries, mind maps, slides, plans, and prototypes
AFFiNE AI is presented as more than a text-generation sidebar. The project highlights workflows such as turning an outline into slides, summarizing an article into a mind map, organizing tasks and backlogs, and generating prototypes or web pages from prompts on the canvas. Because documents and the canvas live in the same workspace, AI output can become part of the working artifact instead of remaining isolated in a chat transcript. Teams using AI should still evaluate the data path of the selected AI service separately from AFFiNE's local-first storage model.
Sources: [1]
For teams combining knowledge management and visual thinking
AFFiNE fits product planning, research, system design, teaching, and personal knowledge workflows where structured writing and spatial thinking repeatedly alternate. It is especially attractive when a team wants page/database organization and a freeform canvas without putting its knowledge base exclusively in a managed SaaS. It is less likely to replace specialized spreadsheet, project-management, or professional design software feature-for-feature; in those cases it works better as the shared knowledge and ideation layer around those tools.
Self-hosting runs a sync backend, and repository licensing is mixed
The official self-host Compose stack runs the AFFiNE server together with Redis, pgvector/PostgreSQL 16, and a migration job, while persisting database, storage, and configuration data. A desktop-only local-first setup therefore has a very different operational footprint from a self-hosted collaborative deployment. Public deployments also need normal service operations around backups, migrations, TLS, authentication, mail, and upgrades.
Licensing deserves a file-level check. The README describes the Community Edition as MIT, but the current root LICENSE carves out packages/backend and packages/common/native under a separate backend license. That backend license says Enterprise Edition portions require a valid subscription while portions distributed as Community Edition are under MPL-2.0; code outside the root carve-outs is available under MIT. The repository should therefore not be treated as uniformly MIT when evaluating redistribution, derivative development, or edition-specific self-hosting.
Official sources
- [1]toeverything/AFFiNE — README(2026-09-14)
- [2]AFFiNE — self-host Docker Compose(2026-09-14)
- [3]toeverything/AFFiNE — root LICENSE(2026-09-14)
- [4]AFFiNE backend — edition license(2026-09-14)
Growth
Growth trends · Last 30 days
0 Stars
Trend data is still being collected.
GitHub data
GitHub dataView detailed GitHub data
- Stars
- 0
- Forks
- 0
- Watchers
- 0
- Open issues
- 0
Report incorrect information
Tell us if any listing information is incorrect or outdated.