Oyma vs Heptabase

Heptabase and Oyma represent two powerful, specialized approaches to personal knowledge management. Both reject the superficiality of traditional linear word processors, and both are designed for serious thinkers who grapple with large volumes of complex information.

However, their core metaphors and architectural foundations are completely distinct.

Heptabase is a visual-first workspace designed specifically for deep qualitative research, learning, and academic synthesis. It approaches knowledge spatially: ideas are atomic “cards” that can be arranged, clustered, and connected across infinite two-dimensional whiteboards. By pairing visual mind mapping with split-screen PDF reading and excerpt highlighting, Heptabase empowers researchers to perceive the underlying structure of complex topics.

Oyma, by contrast, is a focused writing and meeting workspace built around the permanence of plain Markdown text. Instead of encapsulating your ideas in proprietary visual cards, Oyma stores every document as an ordinary .md file directly on your Mac’s filesystem, integrating seamlessly with Obsidian vaults. At the same time, Oyma solves the spoken dimension of modern work: it silently captures video calls without bots, transcribes speech 100% on-device using Whisper, and automatically generates structured meeting records that land straight into your local vault.

Choosing between them depends on whether your intellectual work is primarily visual and spatial, or document-centric and meeting-driven.

Conceptual model: Infinite visual whiteboards vs plain Markdown files

The core philosophical divide between Heptabase and Oyma is how ideas are organized and stored:

In Heptabase:

• Knowledge is organized Spatially. An idea is an atomic “card” that can live on multiple visual whiteboards simultaneously.
• Whiteboards allow you to zoom in and out, draw connection lines, group cards into sections, create mind maps, and annotate PDF documents side by side.
• Data is stored in Heptabase’s proprietary database format, synchronized through its cloud infrastructure with local desktop caching.
• While you can export your cards to Markdown, your spatial whiteboard layouts—the visual groupings, spatial arrangements, and connection lines—cannot be cleanly translated into plain text.
• If you stop subscribing to Heptabase, you lose access to the visual environment that gives your research its spatial meaning.

In Oyma:

• Knowledge is organized Document-Centric. Every note is an ordinary .md text file stored directly on your computer’s filesystem in a folder you choose.
• There is no proprietary database or visual canvas lock-in. A note named Literature Review exists on disk as Literature Review.md.
• You can inspect, edit, or copy your files simultaneously in Obsidian, Visual Studio Code, iA Writer, or standard terminal text editors.
• Links use standard [[WikiLinks]] and metadata uses universal YAML frontmatter. Your thoughts remain permanent, portable text files that can be read on any computer decades into the future.

If your brain thrives on spatial reasoning, concept mapping, and visual paper synthesis, Heptabase’s whiteboard canvas is extraordinarily powerful. If you prioritize absolute plain-text longevity, file ownership, and zero software lock-in, Oyma‘s file-first model is superior.

Meeting capture and spoken audio

The two applications diverge completely when it comes to capturing live spoken conversations:

In Heptabase:

• Heptabase is built for solitary research, reading, and conceptual mapping.
• It contains no meeting detection, no system sound recording, and no automated call transcription.
• On mobile, Heptabase allows quick voice recording for personal memos, and desktop transcription requires supplying your own OpenAI API key.
• To take meeting notes in Heptabase, you must type manually during calls or record conversations with external tools and paste the text into cards.

In Oyma:

• Meeting recording is built directly into the core macOS architecture. On macOS 14.2+, Oyma uses CoreAudio tap APIs to capture both system sound and microphone audio simultaneously without sending an intrusive bot to your call.
• Speech transcription executes 100% on-device using quantized OpenAI Whisper models running on the Apple Silicon Neural Engine. Spoken audio never touches the cloud.
• Oyma transcribes speech in 13 languages, allowing multilingual professionals to capture foreign-language calls without sending confidential audio to remote cloud transcribers.
• Structured AI summaries, action items, and searchable verbatim transcripts are saved automatically as Markdown files directly into your vault, cross-linked to your daily notes.

For knowledge workers whose weeks are filled with video calls, design syncs, and client briefings, Oyma bridges the divide between spoken conversation and written documentation without compromising privacy.

Artificial intelligence: Credit-metered tiers vs sovereign local models

How artificial intelligence is delivered reflects each application’s business model:

In Heptabase:

• Heptabase includes an AI assistant with chat and “Agent Mode” to query cards and summarize documents.
• However, AI access is restricted by subscription tiers: the $8.99/month Pro plan is limited to Gemini models on monthly credits, while frontier models from OpenAI and Anthropic (Claude) require upgrading to the $17.99/month Premium tier.
• Context and queries are sent to cloud AI providers via Heptabase’s servers.

In Oyma:

• AI in Oyma is sovereign, transparent, and user-directed.
• You can run open-weight language models (such as Llama 3.2, Mistral, or Qwen 2.5) locally on your Mac via native Ollama integration. Processing runs entirely on your local GPU and Neural Engine with zero data leaving your machine.
• If you prefer frontier reasoning models, you can supply your own API keys (Bring Your Own Key) for Anthropic or OpenAI. Requests travel directly over TLS 1.3 socket connections from your Mac to the model provider, completely bypassing intermediary proxies.
• A built-in egress log provides an auditable real-time record of all network activity, proving that your private notes and meeting transcripts remain strictly on your device.

Platform support and academic workflows

The intended use cases highlight distinct operational advantages for each tool:

Where Heptabase wins:

• Academic and qualitative research: The built-in PDF reader with split-screen card extraction, visual mind mapping, and infinite whiteboards makes Heptabase one of the best tools in existence for synthesizing complex academic literature.
• Cross-platform availability: Native applications across macOS, Windows, Linux, iOS, Android, and web ensure access to your research anywhere.
• Developer extensibility: Includes a CLI and MCP (Model Context Protocol) integration for external agents.

Where Oyma wins:

• Native Apple Silicon performance: Built specifically for macOS, Oyma launches in milliseconds, handles massive text files without lag, and conserves battery life.
• Direct Obsidian vault compatibility: Point Oyma at an existing Obsidian vault, and both applications work harmoniously on the exact same files without friction or export routines. (Note that Obsidian also includes an infinite visual Canvas for users who want visual mapping).
• Spreadsheet-style tables: Inline Markdown tables feature column sorting, keyboard navigation, and alignment controls while remaining standard GitHub Flavored Markdown on disk.
• Zero subscription overhead: Free during private beta for Apple Silicon Macs, with no recurring monthly subscription fees ($8.99–$17.99/mo in Heptabase).

Pricing and ownership comparison

Heptabase operates on a paid-only subscription model:

• No permanent free tier: Users must subscribe to access the software.
• Pro: $8.99 per month billed annually ($107.88/year) or ~$11.99 monthly, including Gemini-only AI credits.
• Premium: $17.99 per month billed annually ($215.88/year), adding Claude and GPT frontier models.

Oyma operates on a sovereign hardware model:

• Free during beta: Full access to all features, including meeting recording, Whisper transcription, and local AI.
• Zero server costs: Audio transcription and local AI run on your Mac’s hardware.
• Wholesale cloud AI: If you use cloud models, you pay token costs directly to the model provider with zero markup.
• Permanent files: If you discontinue using Oyma, your notes remain as standard .md files in your Finder folders.

The verdict

Choose Heptabase if:

• You are a researcher, academic, or student who needs infinite visual whiteboards to organize atomic concepts and mind maps.
• You do intensive PDF reading and want to extract highlights directly into visual cards.
• You need cross-platform applications across Windows, Linux, Android, iOS, and web.
• You are comfortable paying $8.99 to $17.99 per month for a specialized visual thinking environment.

Choose Oyma if:

• You insist that your notes remain plain Markdown files stored on your local disk with zero vendor lock-in.
• You participate in video calls and need private, on-device Whisper transcription that captures meetings without bots.
• You already use or plan to use Obsidian, and want meeting notes and daily records to land directly inside your existing vault.
• You want unlimited offline AI via Ollama or transparent BYOK cloud inference without monthly subscription paywalls.

Where Heptabase is better

• Visual whiteboards for spatial thinking and research synthesis.
• Apps on every platform including Windows, Linux, iOS, Android and web.
• CLI and MCP access for external agents.

Thinking of moving? Your notes in Oyma are plain Markdown, so you can always take them with you. Start withgetting started, or read what leaves your Mac.

Looking at other options too? Read our guide to the best Heptabase alternatives.