SumizAI vs AFFiNE

What AFFiNE is

What AFFiNE is

AFFiNE is an open-source, privacy-focused workspace that combines note-taking, project and task management, a Kanban board, an interactive whiteboard, and a Pomodoro timer into a single app. Built in Singapore, it quickly earned recognition in the open-source community for an ambitious premise: instead of forcing users to switch between separate apps for notes, tasks, and visual planning, AFFiNE merges all those work modes into one coherent document that can be freely toggled between a text view and a visual canvas.

Key features

AFFiNE's foundation is a block-based editor concept, where every element — a paragraph, a list, a table, an embedded file — is an independent block that can be moved, nested, and formatted independently of the rest of the document. The same document can be viewed as linear text in a Markdown-style editor or as nodes on an infinite whiteboard, with no need to copy content between two separate tools. A Kanban board and goal tracking let you manage project tasks right alongside notes about the same topic, eliminating information fragmentation across different apps. Real-time team collaboration with @mentions lets multiple people work on the same document simultaneously. Offline access and a self-hosting option give full control over where data is physically stored — on your own server or locally on device, with no dependence on the vendor's cloud.

Who it's for

AFFiNE appeals to teams and individuals looking for one tool instead of a stack of separate apps — a notebook, a task manager, a Kanban board, and a visual planning tool. People who value open source code and the ability to self-host their data, skeptical of closed SaaS platforms, find in AFFiNE a rare combination of rich functionality with full infrastructure control. Project teams that combine visual planning with text documentation — product or creative teams, for instance — can benefit from switching the view of the same document between text mode and a board.

Use cases

In practice, AFFiNE works well for sprint planning, where a team starts from a Kanban board of tasks, but any task can expand into a full text note with context and decisions. A UX designer can use the whiteboard to map a user journey, then switch to text view to write up research findings in the same document. A freelancer managing several client projects at once uses the Pomodoro timer to track work time right alongside notes for a specific engagement. A geographically distributed team coordinates work through @mentions in comments, avoiding scattered communication across separate chat channels.

Pricing and business model

AFFiNE runs on a freemium model — the core app is free and open source, with self-hosting available at no cost. Paid plans mainly cover the cloud service hosted by the creators, offering convenient sync without managing your own server, extra storage, and priority support. That model lets technical users tap the full functionality for free while offering a convenient paid alternative for people who don't want to handle their own hosting.

Limitations and what to watch for

The wealth of work modes — text, whiteboard, Kanban, goals — can feel overwhelming for new users looking for a simple notebook rather than an elaborate productivity platform. Self-hosting requires basic technical knowledge, so people unwilling to deal with servers will depend on the paid cloud service. As a relatively young open-source project, AFFiNE is still actively evolving its features, which means possible changes to the interface and data structure between releases.

AI notes in AFFiNE

AFFiNE has no built-in native AI model integration for automatically generating content from conversations — if a user wants to keep an ai notes entry from a session with an external AI assistant, they have to paste it in manually as a new block in the document, using the flexible block structure to place it alongside related Kanban tasks or text notes. The block structure makes it easy to later connect that pasted ai notes entry with the rest of a project's context, but the pasting process itself remains manual.

Migrating from other tools

Teams moving to AFFiNE from a scattered toolset (a separate notebook, a separate Trello, a separate Miro board) often report that the biggest benefit isn't the functionality itself but the reduction in context-switching — fewer browser tabs, fewer notifications from different apps, one place to search when you need to find a specific project decision. Importing existing notes from other Markdown formats usually goes smoothly, though more complex structures (like nested Notion databases) may need manual cleanup after import. People revisiting old ai notes pasted in manually from AI conversations can freely move them between text blocks and the visual board, depending on how they're currently organizing a given project.

What's worth checking before choosing

Before fully switching to AFFiNE, it's worth testing both work modes — text and board — on the same document, to judge whether the flexibility of switching views actually fits your working style or introduces unnecessary complexity. It's also worth checking whether self-hosting is realistic given your available technical resources, or whether it's better to budget for the paid cloud service from the start.

Bottom line

AFFiNE is an ambitious, open-source attempt to combine a notebook, task manager, and visual board into one tool, with full control over data thanks to self-hosting. Anyone who wants ai notes from AI model conversations to land automatically in such an integrated system without manual pasting should consider a tool built specifically around that workflow.

Comparison with SumizAI

AFFiNE and SumizAI share a philosophy of openness and user control over data, but differ in scope. AFFiNE is a broad productivity platform combining multiple work modes — notes, tasks, a visual board — built up by hand by the user. SumizAI focuses narrowly on one thing: turning a conversation with an AI model into an ai notes entry saved as a Markdown file in a vault the user owns, with no manual pasting. Anyone who needs a full project management platform will pick AFFiNE; anyone who wants ai notes from AI conversations to land automatically in a simple, portable file archive will find SumizAI a better-fitting, narrower tool.

That is our reading of the record. For AlternativeTo's own wording, see its page for AFFiNE — the facts above were read on 2026-08-16.

AFFiNE in depth

One canvas that is both a document and a whiteboard

AFFiNE's structural idea is that a page has two modes rather than two homes. The same content can be viewed as a linear document — the Page mode, with block-based rich text — or opened in Edgeless mode, an infinite canvas where those blocks become objects you can move, connect and draw around. There is no export step between the two and no separate whiteboard file; a meeting note and the diagram sketched during that meeting are the same object seen from different angles. On top of that sit databases with kanban and table views, templates, and an AI layer that is sold separately.

The project is developed by TOEVERYTHING PTE. LTD., a Singapore company, and it is one of the largest open-source workspaces by community size: the GitHub repository was created in July 2022 and now carries roughly 71,700 stars and 5,100 forks, with active development — the canary branch has more than eleven thousand commits and the repository receives pushes daily. It is written in TypeScript, with an Electron desktop shell and React interface. Underneath it uses BlockSuite as the collaborative editor framework, Yjs for conflict-free replicated data types, and OctoBase, a Rust data layer, which is why concurrent edits merge rather than conflict.

The licence is dual, and the distinction is the important part

AFFiNE is frequently listed as simply open source, and just as frequently as proprietary. Both labels are wrong, and the vendor is unusually clear about why. The repository root licence file states that content under packages/backend and packages/common/native is licensed under a separate licence, while everything else is available under MIT. That separate licence is the AFFiNE Enterprise Edition licence, and its terms are strict: production use requires a valid AFFiNE Enterprise Edition subscription for the correct number of seats, modifications and patches remain the vendor's property, and copying, merging, publishing, distributing, sublicensing or selling the software is forbidden outside those terms. Development and testing copies are exempt.

The pricing FAQ restates this in plain language: the editor, the desktop application and most of the codebase are MIT-licensed, while the self-hosted backend is source-available under the EE licence, and without a commercial agreement you may read the backend source only — modifying, compiling or redistributing it requires authorisation. Anyone planning a self-hosted deployment should read that paragraph before anything else on the site, because it is the difference between a permissive project and a commercially gated server.

Plans and what each one actually caps

The free tier is generous on the local side and metered on the cloud side. Locally you get unlimited workspaces, unlimited customisation and unlimited document and edgeless editing. The bundled Cloud Basic account adds 10 GB of cloud storage, a 10 MB maximum file size, up to three members per workspace, seven days of what the vendor calls Cloud Time Machine version history, and up to three logged-in devices. There is no time limit.

Pro costs 6.75 US dollars per month and raises those numbers to 100 GB of storage, a 100 MB maximum file size, ten members per workspace and thirty days of version history. Team is 10 dollars per seat per month, starting at ten seats, with 100 GB of initial storage plus 20 GB per seat, a 500 MB file ceiling, unlimited members, multiple admin roles and priority support; the same tier exists for self-hosted deployments. There is a one-off Believer plan at 499.99 dollars giving lifetime personal use of Pro with 1 TB of storage. Enterprise pricing is custom, as is a separate commercial code licence covering OEM use, white-labelling and embedding AFFiNE in another product.

AI is not included in any of the above. The AFFiNE AI add-on is priced at 8.9 dollars per month billed annually and covers writing and rewriting, turning outlines into presentations, summarising content into mind maps, translation, grammar fixes and action-item extraction; auto-sorting and auto-tagging of notes are listed as still to come. The vendor states 200,000-plus Pro users on the pricing page, which is a self-reported figure.

Platforms, self-hosting and data ownership

Desktop builds exist for macOS, Windows and Linux, alongside a web client and mobile applications; the client keeps a local copy and syncs to the cloud when a workspace is cloud-enabled, so offline editing is a normal state rather than a degraded one. Self-hosting runs the full stack on your own infrastructure, where the only quantitative limits are your own hardware — but as above, that path is governed by the EE licence rather than by MIT.

The local-first design has a practical consequence worth stating: a purely local workspace has no server dependency at all, which means a subscription lapse or a service outage does not lock you out of documents that were never in the cloud. It also means that the strongest data-ownership story in AFFiNE is the one where you use the least of the paid product.

Known trade-offs

The scope is very large for a single application, and the issue tracker reflects it — several hundred open issues at any time is normal for a project of this size, but it also means feature depth varies. The databases are not a spreadsheet replacement, the edgeless canvas is not a full design tool, and the AI features are add-on grade rather than core. Users evaluating it against Notion should expect a smaller template and integration ecosystem; users evaluating it against Miro should expect a whiteboard that is a mode of a document editor rather than a purpose-built canvas.

The licence split is the other recurring friction point, because it surprises people who arrive expecting an MIT project and discover that the component they wanted to run on their own server is the one that is gated. That is a legitimate commercial model, clearly disclosed, but it should be a conscious choice rather than a discovery made after deployment.

Who it fits

AFFiNE suits individuals and small teams who want documents and visual thinking in one tool, who value local-first storage, and who are comfortable with a fast-moving project. It suits organisations with a self-hosting requirement only if they are prepared to engage commercially. It fits poorly where a mature integration catalogue, formal support guarantees or a stable long-term feature surface are the priority. SumizAI overlaps only at the edges — it processes and summarises documents rather than providing a workspace to author them in.

Product page: affine.pro

What SumizAI is

A note-taking application built around a conversation with an AI model. You ask a question, the answer streams back, and the answers worth keeping become Markdown notes — filed into a vault that is a folder on your own disk.

A vault is a directory holding base.md, a generated table of contents up to six levels deep, and a notes/ folder with one .md file per note. Before a note is written it is checked against the ones already there, so the fourth note about the same idea gets merged instead of added. Links between notes are ordinary Markdown links to files that exist.

The model is never ours: you bring your own API key to one of seven providers — Anthropic, OpenAI, Gemini, Groq, OpenRouter, a local Ollama or your own server — and pay that provider directly. A question sends the table of contents plus at most five relevant notes within a 24,000-character budget, and the app shows you which five it used.

Every answer kept this way becomes one of these ai notes, ready to search again months later without hunting back through old chats.

Side by side

Where they differ

• AlternativeTo's description of AFFiNE does not mention an AI model. That is the whole starting point in SumizAI: you have a conversation with a model on your own API key, and the answers worth keeping become notes without you filing them.
• AFFiNE is freemium: there is a free tier and a paid one, and the line between them is the thing to read before committing. SumizAI has one plan at a dollar a month, so there is no feature held back for a higher tier — and an unpaid account still opens and exports every vault it already has.
• Both keep Markdown, which makes moving between them mostly a matter of copying files. In SumizAI a vault is a directory holding base.md — a generated table of contents — and a notes/ folder with one .md file per note.
• AFFiNE is described as something more than one person uses at once. SumizAI is not: there are no shared vaults, no comments, no permissions and no sync between devices. If the work is a team's, that is a reason to pick AFFiNE over SumizAI.
• AFFiNE can be self-hosted. SumizAI has nothing to host on the desktop — the vault is a folder on your machine — and if you want the model on your own hardware too, a local Ollama is one of the seven providers it speaks to.
• AFFiNE also does tasks. SumizAI does not — no boards, no due dates, no projects. It files knowledge, not work.
• AFFiNE offers a visual surface — a canvas, whiteboard or map. SumizAI has no canvas: the structure it maintains is a written table of contents in base.md, up to six levels deep, rebuilt as notes arrive.

Which one to pick