Comparison 03 · Researched Oct 04, 2026

AFFiNE vs SiYuan

Two serious open knowledge tools, two very different personalities. AFFiNE wants to replace your docs + whiteboard + workspace stack. SiYuan wants to give power users a deep block-based knowledge system they truly control.

AFFiNE

Docs · Edgeless canvas · databases · collaboration

Visual workspace

Product briefWrite the launch story

Idea mapConnect concepts freely

DatabaseStatus · Owner · Date

Team noteComment + collaborate

SiYuan

Block references · SQL · backlinks · self-hosted PKM

Power-user PKM

Atomic blockref
Every idea can become an addressable unit.

Linked contextbacklink
References stay connected across documents.

Query your vaultSQL
Surface knowledge by structure, not folders alone.

Reuseembed
Pull the same block into multiple contexts.

Why this keyword belongs on the site

The exact comparison SERP is still thin. One recent dedicated self-hosted comparison exists, plus aggregator coverage, but there is nowhere near the saturation you see around Notion or Obsidian. Meanwhile both projects have large, active GitHub communities and fit the same high-intent “open-source / self-hosted / privacy-first knowledge base” search cluster.

SEO angleLow exact-match crowding, broader intent underneath.AFFiNE vs SiYuan → open-source Notion alternative → self-hosted notes → local-first PKM → whiteboard vs block references

Fast verdict

Choose the interface philosophy first.

AFFiNE is easier to understand if you think in pages, whiteboards and projects. SiYuan becomes more powerful if you think in blocks, references, queries and reusable knowledge.

Best for visual workspace users

Choose AFFiNE when you want one place for writing, databases and freeform visual thinking.

Its strongest differentiator is the combination of conventional docs with an “Edgeless” canvas and workspace-style collaboration. AFFiNE Community Edition is currently available for self-hosting under MIT terms.

Visual workspace

9.7

Team familiarity

9.1

PKM depth

7.8

Best for block-level knowledge

Choose SiYuan when the smallest useful unit in your knowledge base is a block, not a page.

Fine-grained references, two-way links, custom attributes, SQL query embeds, databases, flashcards, web clipping and Docker deployment make it unusually deep for technical and long-term personal knowledge systems.

Block-level PKM

9.9

Power-user depth

9.6

Visual freedom

6.9

What are you trying to build?

For a study system, SiYuan gets the edge if you value backlinks, block references, PDF annotation and built-in flashcard/spaced-repetition workflows.

Side by side

Where the products really separate.

Both can be local and self-hosted. The bigger question is what happens after your data is inside.

Decision	AFFiNE	SiYuan	Edge
Core mental model	Docs + databases + infinite/Edgeless canvas	Documents made from addressable content blocks	Preference
Visual whiteboarding	A core product feature; visual work and docs live together	Supports diagrams/charts, but visual canvas is not the central model	AFFiNE
Block references	Flexible block editor and linking	Fine-grained block reference, embedding and two-way links are core strengths	SiYuan
Advanced querying	Databases and workspace views	Custom attributes plus SQL query embeds for deep retrieval	SiYuan
Self-hosting	Official Docker/self-host path; Community Edition available	Docker deployment and local-first/self-hosted workflows are core use cases	Both strong
Licensing	Community Edition currently MIT; enterprise/backend licensing differs by edition	Main SiYuan repository is AGPL-3.0	Different models
Study features	Strong documents, visual mapping and AI-assisted knowledge workflows	PDF annotation linking, flashcards/spaced repetition, web clipping	SiYuan
Team / workspace feel	Closer to modern collaborative workspace products	Primarily feels like deep personal knowledge software	AFFiNE

Workflow diagram

Same knowledge. Different architecture.

Imagine researching a new topic. The two apps encourage very different paths from raw information to understanding.

AFFiNE → spatial workspace

01 · Capture notes in documents

02 · Put concepts onto Edgeless canvas

03 · Group ideas visually

04 · Add structured database/project data

05 · Share and collaborate in one workspace

SiYuan → atomic knowledge

01 · Capture information as blocks

02 · Reference exact blocks elsewhere

03 · Build backlinks and reusable context

04 · Query attributes / databases / SQL

05 · Re-surface knowledge through links and review

Community intent

The long-tail opportunity is bigger than the exact keyword.

People in PKM and self-hosted communities repeatedly ask for the same things: a real open-source Notion alternative, local data, less dependence on cloud vendors, strong databases, and tools that do not force one style of thinking.

Search intent 01

“I want a real open-source Notion alternative I can keep local.”

SiYuan is actively recommended in that exact kind of 2026 PKM discussion.

Search intent 02

“Can I get docs, whiteboards and databases without using Notion + Miro?”

That is almost exactly the market position AFFiNE is pursuing.

Search intent 03

“Which self-hosted knowledge tool will still work when my setup gets complicated?”

This lets us target future pages around Docker, sync, licensing, block references, web clipping and AI.

Final call

AFFiNE for the workspace. SiYuan for the knowledge engine.

If your ideal app feels like Notion and Miro merged into one local/self-hostable workspace, start with AFFiNE. If you want granular block references, powerful retrieval and a serious long-term PKM system, start with SiYuan.

Choose AFFiNEVisual planning · docs · canvas · collaborative workspace · familiar structure

Choose SiYuanBlock references · backlinks · SQL · flashcards · deep self-hosted PKM

Research checked Oct 4, 2026. Product features and licensing can change; verify official pages before making deployment decisions.
Sources: AFFiNE official · AFFiNE GitHub · AFFiNE enterprise/self-host · SiYuan GitHub · SelfHostIndex comparison · LibHunt comparison · Reddit PKM discussion.