Acres is our first recommendation for a personal visual note workspace on Mac or Windows. The case is practical: your cards are Markdown files, your boards are readable JSON, and $59 buys the app once at the current early-adopter price. You can arrange research by hand, ask the built-in assistant to help, or connect an agent over MCP.
Heptabase is the more developed research environment. Its broader device support and learning tools can justify a subscription. We make Acres, so this recommendation reflects the priorities behind our own product: files you can keep, a canvas you use every day, and AI you can choose. It is not a claim that Acres has every Heptabase feature.
Acres Plus is also available for $9/month, including app access, Sync, and Publish with no upfront purchase. The $59 lifetime option covers local use.
What does each app cost over a year?
Heptabase Pro is $8.99 a month on annual billing, or $107.88 for a year. Acres is $59 once. For the core app, that puts Acres $48.88 below one year of Heptabase Pro, before taxes and promotions. The gap grows over time if those prices remain the same.
That calculation has a boundary. Heptabase bundles cloud services into its plans; Acres sells optional Sync and Publish separately. Hosted AI usage with your own key is billed by the provider. Compare the services you actually need before treating the license difference as your total saving.
The early-adopter price also reflects where Acres is today. It’s a newer, less mature app, with fewer integrations and more rough edges than Heptabase. The changelog shows what is improving: recent releases added full-text command-palette search, clearer imports and LM Studio support. Buy for the features available now.
A database or a folder of notes
Both apps let you arrange cards, connections and related material on whiteboards. Both support offline work. The distinction appears when another tool needs the notes.
Heptabase stores working content in its app database. Its official guidance tells agents to use the CLI instead of manipulating internal storage. That’s sensible: writing into a database behind an app’s back can damage it.
Acres stores a card in a file such as cards/interview-notes.md. The board stores that card’s position in JSON. A note can appear on several boards without creating several copies of its text. You can back up the folder, read a card in another editor or let an agent inspect it with Acres closed. Acres links use its own acres:// addresses, and its board schema is specific to Acres; plain files don’t mean every app understands every relationship.
For someone building a long-lived personal archive, that access is why we’d start with Acres.
AI: both have MCP, but the workflows differ
Heptabase deserves full credit here. Its MCP server supports external assistants, including keyword and semantic search. Its desktop CLI can edit cards and manipulate whiteboard layouts, sections, connections and mind maps. It also accepts your own API keys. BYOK is not exclusive to Acres.
Acres brings agent tools into a local file workspace. MCP is bundled with the app, and Settings can connect supported clients such as Claude Code, Cursor and Codex. Agents can find notes, create cards and arrange material on a board. Activity history helps you inspect supported changes and revert them with safeguards against overwriting later work.
Inside Acres, card rewrites and larger layout changes arrive for review. Some additive actions apply immediately. Choose a cloud provider with your key, or run supported models through Ollama or LM Studio locally. Local AI still needs a capable machine and a model that can follow the tool workflow. The local MCP guide explains setup and review steps.
For people who want agents working in files they own, we prefer Acres’ combination. For guided study and hosted research assistance, Heptabase has capabilities Acres doesn’t offer.
When Heptabase is the better choice
A researcher who reads on an iPad, captures ideas on a phone and meets colleagues on a shared board needs more than a good desktop canvas. Heptabase covers that routine. Its pricing page includes PDF parsing and OCR on higher plans, an AI Tutor and bundled AI credits.
Acres has a PDF reader, highlights and linked cards. It has no mobile app, Linux build or real-time collaboration, and it doesn’t replace Heptabase’s full research stack. If one of those gaps blocks your work, Heptabase should win your decision. The Heptabase, Noteey and Acres guide also covers the middle option: Noteey’s local multimedia workspace.
Trying Acres with an existing Heptabase library
Use the built-in importer from the Acres launcher. Take a full manual Heptabase backup, keep the original archive, then import a copy. Inspect the warnings and check a few demanding examples: a nested board, a PDF with highlights and a card with links or formatting.
The import preserves board positions, which saves rebuilding your maps from a flat Markdown export. It is a snapshot, not a live bridge between the apps. Future Heptabase edits won’t appear automatically in Acres.
Use the 14-day trial for a real project. You should be able to find an old note, arrange its context and get useful work from an agent before deciding to move. For a broader shortlist, see our infinite canvas app guide.