Acres is our first recommendation for a personal canvas you can buy once and keep as local files. At $59 during the early-adopter period, it gives you Markdown notes, a visual workspace, a built-in assistant and a bundled MCP connection. It is a strong place to start if your work lives on a Mac or Windows computer and you want an agent to help organize it.
We make Acres. This recommendation weights file access, personal canvas work, AI control and app cost. Heptabase and Noteey have advantages that can outweigh those priorities. Heptabase’s connected reading workflow and mobile apps are substantial; Noteey’s timestamped media notes and presentation tools solve jobs Acres doesn’t currently match. Neither should be dismissed as a more expensive version of the same product.
The buying decision in one table
| Dimension | Acres — our first pick | Heptabase | Noteey |
|---|---|---|---|
| Main reason to choose | Local-files canvas with external agents | Integrated research across devices | PDF, video and audio study on a canvas |
| Paid starting point | $59 once, early-adopter license | Pro $8.99/month billed yearly | Desktop Pro $99 lifetime |
| Free option | 14-day trial | Trial; no ongoing free plan | Three projects and 100 cards |
| Working storage | Markdown cards and readable JSON boards | Local desktop database; cloud services when enabled | Local projects; portable project copies |
| Research emphasis | PDF highlights, cards, connections and tasks | Reading integrations, annotation and AI tutoring | Media timestamps, annotations and floating reading windows |
| External agent access | Bundled local MCP | Official remote MCP | No official MCP setup found in the sources reviewed |
| AI model choice | Your provider key, Ollama or LM Studio | Included credits and documented BYOK options | Premium summaries and transcripts |
| Important limit | Newer desktop app; no mobile or real-time team editing | Recurring app cost | Desktop/Premium plan boundaries; test exports |
This is a comparison of published features and prices, not a claim that we benchmarked every workflow. The scope of each plan matters more than a check mark.
Compare the bill you will actually pay
For the base paid app, the current prices are:
| App | First year | Three-year illustration at today’s base price |
|---|---|---|
| Acres | $59 | $59 |
| Noteey Desktop Pro | $99 | $99 |
| Heptabase Pro, annual billing | $107.88 | $323.64 |
The Heptabase figures are $8.99 multiplied by 12 and 36 months. The one-time figures assume you keep the purchased license. This illustration excludes tax, optional services, AI usage, promotions and future price changes; it isn’t a promise of feature-equivalent total ownership cost. Sources: Acres pricing, Noteey pricing and Heptabase pricing.
Acres is $40 below Noteey Desktop Pro’s list price and $48.88 below the first year of Heptabase Pro. Those savings make it worth trying, but a missing feature can matter more than the difference.
Noteey’s Premium subscription is a separate comparison: $10 monthly or $90 yearly, with a $599 lifetime option listed at the time of checking. It bundles services such as AI summaries and transcripts. Desktop Pro alone doesn’t represent that whole offer. Noteey also advertises promotions, so check the checkout price. Its free tier can be enough for a small project.
Acres’ assistant and MCP tools are included with the license; cloud models are billed by your provider or external agent service. Optional Acres cloud plans have separate charges. A local Ollama or LM Studio workflow changes that calculation again, but requires suitable hardware and a model that can do the task.
If your day starts with reading
Choose Acres first when you mainly read PDFs, collect written notes and build an argument on your desktop. Highlights can become material on the board, and cards can hold evidence, questions and conclusions beside the source. Tasks can stay in those cards instead of moving to another system.
Choose Heptabase when your research depends on Readwise, Zotero, a web clipper or continuing on your phone. Its research feature overview shows how those sources fit into a connected knowledge base. The integration saves work that a lower app price won’t necessarily recover.
Choose Noteey when the source is a lecture, interview or video course. Its video notes workflow ties notes to timestamps; its PDF tools support annotations placed into visual work. Presentation mode is another reason to choose it when you will show the board to an audience.
Local storage and an exit path are separate checks
Acres keeps cards as Markdown and board structure as JSON in a folder you choose. That makes the writing accessible to editors and scripts without requiring an export. Another app won’t automatically reproduce every Acres canvas feature, but it can read the note text.
Heptabase’s desktop app stores a local database for offline use. Cloud sync and AI introduce additional processing; its data handling documentation explains the distinction. It also documents what remains accessible when a subscription ends. Check that process as part of the trial.
Noteey’s backup guide describes portable .pcanvasx projects and full backups. Its changelog also documents Markdown export for cards. That is a useful exit route, even though it differs from using Markdown as the working format.
Before moving a library, test two linked notes, one PDF and one board. Open the resulting text or project elsewhere. Check the attachments and links as well as the words.
AI and MCP: compare the work, not the badge
Acres is especially useful when an outside agent needs to search your notes and work on their visual arrangement. It can propose a board layout, show a preview and apply the approved result. The built-in assistant offers another route with your selected provider or local model. The agent guide explains review and supported revert.
Heptabase also has a capable official MCP service, with read/write authorization and keyword or semantic search. It also supports your own API keys. Claims that it only has metered built-in AI miss those options.
Noteey’s documented AI emphasis is summaries and transcripts, including local media summaries described in its release history. We didn’t find an official MCP setup in the reviewed product, pricing or help pages. If external agent access is essential, ask the vendor about current support.
Our preference for Acres is about where the agent works and how the result fits the board. We haven’t measured model answer quality across these apps and don’t claim universal AI superiority.
The maturity trade-off is real
Acres is a young app. Its release history includes improvements to imports, content search, drawing and local AI support. Progress is useful evidence, but it doesn’t replace an established workflow or make a promised feature available today.
Try a small project before committing. Check editing, reopening the board, import warnings and a representative AI task. If those steps feel right and your work is personal desktop research, Acres’ combination is the one we recommend first. If mobile research integrations or timestamped study material are essential, Heptabase or Noteey may justify their cost. For the broader field, read our infinite canvas shortlist.