Heptabase vs.
Innogath.

Two opposite directions on the curation axis

The clearest way to understand the difference is on a single axis: who creates the structure?

Heptabase is manual. You open a whiteboard, write cards, drag them around, connect them. The structure is whatever you make it. The accuracy of the synthesis equals the accuracy of your reading. The freedom of the layout is total — free-form spatial canvas, mind maps, sections, drawn arrows. After two years of careful curation, your Heptabase looks like a personal knowledge garden you grew yourself, and that’s exactly what serious users love about it.

Innogath is AI-first. You ask a question. An agent fans out across the open web, fetches 20-50 sources, writes a multi-chapter cited report, and arranges follow-ups as a branching tree. The structure is generated; you edit and refine. The accuracy of the synthesis depends on the agent and the sources it found, with confidence scores surfaced inline. The freedom of the layout is limited — tree-shaped, not free-form — but the speed of getting from question to defensible deliverable is dramatically faster.

Neither approach is universally better. They serve different stages of intellectual work. Heptabase is for ideas you’ve already had and want to organise. Innogath is for ideas you haven’t had yet. Most serious knowledge workers need both at different points in the same project lifecycle.

When Heptabase wins

Heptabase is the right tool for shapes where the work is the curation:

• Long-running personal knowledge management. Years of reading distilled into cards, connected over time, kept local. This is Heptabase’s home turf and nothing else does it as well. Innogath has no equivalent metaphor for “the knowledge garden you grow over a decade.”
• Visual thinking on a free-form canvas. Drawing arrows, grouping cards spatially, drafting frameworks where the layout itself carries meaning. Innogath’s tree is more rigid by design.
• PDF annotation and highlight workflow. Heptabase’s PDF reader integration — read, highlight, send the highlight to a card — is a genuinely well-built loop. Innogath does not have a comparable PDF annotation surface.
• Local-first storage and offline editing. Heptabase keeps your data on disk; you can work offline, control your file format, back up on your terms. Innogath is cloud-first; if connectivity is critical to your concerns, this is a real difference.
• Native Mac, iOS, and Windows apps. Heptabase ships native clients with all the polish that implies — keyboard shortcuts, menu bars, OS integration. Innogath is a desktop web app today.

If your work is “I have ideas, help me organise them visually over years,” Heptabase is the tool. The manual nature isn’t a limitation — it’s the whole point. The cards are slow to build because you’re building them carefully, and that’s where the value lives.

When Innogath wins

Innogath is the right tool for shapes where the work starts before you have ideas to organise:

• Open-ended research investigations. “What’s the climate tech VC landscape in 2026?” — you don’t have cards yet; you need an agent to find sources, read them, synthesise, and cite. Heptabase can’t do this; you’d have to do all the source-finding and reading manually first.
• Cited deliverables. Multi-chapter reports with structured bibliography, exportable to DOCX with citations preserved through edits. Heptabase exports PDFs of your whiteboard, but doesn’t generate bibliographies or cite sources programmatically.
• Multi-week research with branching follow-ups. Each follow-up question becomes a child page that inherits parent context and pulls in fresh sources. Heptabase has cards but not auto-generated branching from a research question.
• Cross-source verification. Innogath cross-checks claims across many independent sources. Heptabase has whatever sources you’ve manually copied into cards.
• Re-runs over time. Re-run a research question next month — fresh sources, same tree structure, diff against last month. Heptabase requires manual re-research.

In these shapes, the work isn’t curating ideas you already have. It’s bringing back evidence you don’t have yet. That’s a fundamentally different job, and it needs a fundamentally different tool.

Whiteboard freedom vs tree structure

This is the trade-off worth being honest about.

Heptabase’s whiteboard is more flexible than Innogath’s canvas. You can lay out cards in any spatial pattern — by importance, by chronology, by argument structure, by hand-drawn quadrants. The layout itself becomes part of the meaning. For visual thinkers, this freedom is the point of the tool; constraining it would defeat the purpose.

Innogath’s canvas is more structured. The diagrams are auto-generated from your report (22 chart types: comparison matrix, timeline, decision tree, sankey, etc.); you edit them, but you don’t draw arbitrary spatial layouts. The branching tree on the left is hierarchical, not free-form. For users who want the structure to be inferred rather than invented, this is the right design. For users who think visually and want the canvas to carry meaning, it can feel rigid.

The honest answer: if you’re a fundamentally visual thinker who reasons by spatial arrangement, Heptabase’s whiteboard is closer to how you work. Innogath’s auto-structure is good for a different kind of mind — one that wants the tool to take a first pass at organising and then refines from there. Neither is universally correct.

The lifecycle question

A useful frame: Heptabase and Innogath occupy different stages of the same intellectual project lifecycle.

Stage 1 — Discovery. You don’t yet know what’s true. Innogath is the right tool: agentic web research, cited synthesis, branching follow-ups. The output is a structured report you couldn’t have written manually because you didn’t have the sources yet.

Stage 2 — Synthesis into a deliverable. You have the research; you need to write the brief, paper, or memo. Innogath’s notebook with live citations works for this; Heptabase’s whiteboard works if you prefer visual drafting. Either is reasonable.

Stage 3 — Curation into long-term knowledge. You’re done with the project; you want to keep what you learned in a place that compounds over years. Heptabase wins here. The whiteboard becomes a permanent layer of your personal knowledge graph; Innogath’s project tree is good for that project, but not for cross-project knowledge that you’ll come back to in three years.

The handoff between Innogath and Heptabase is exporting from Innogath (Markdown or PDF, with citations preserved) and importing into Heptabase as cards. Several users we talk to run this exact pipeline.

Visual thinker vs structured thinker

Worth naming directly: the two tools reward different kinds of thinking.

Heptabase rewards visual reasoning. If you’re the kind of person who solves problems by drawing on a whiteboard, arranging notes by spatial proximity, or sketching concept maps with arrows pointing different directions — Heptabase is built for you. The whiteboard isn’t decorative; it’s where the thinking actually happens. Constraining it to a tree structure would lose the value.

Innogath rewards structured reasoning. If you’re the kind of person who solves problems by outlining, by hierarchical decomposition, or by working from a research question down through sub-questions — Innogath’s tree-shaped workspace matches. The auto-generated structure is a starting point; you don’t have to invent the outline yourself. Constraining it to a free-form whiteboard would lose the value.

This isn’t a hierarchy of thinking styles — both are valid, and most serious researchers use both at different stages. But it’s a real factor in tool choice, and naming it honestly is more useful than pretending the two tools are interchangeable. Try both before committing to one; the gut feel after thirty minutes of use is usually right.

Data ownership and privacy

A real difference worth knowing if you’re choosing between them.

Heptabase is local-first by design. Your cards live as files on your disk; you control the file format, the backup, and the deletion. Heptabase’s cloud sync is a convenience layer; the source of truth is local. For users with hard data-residency requirements, regulatory constraints, or simply a preference for not putting their long-term knowledge in someone else’s cloud, this is a real and ongoing advantage.

Innogath is cloud-first. Your projects live on our servers (with end-to-end encryption in transit, encryption at rest, and standard SOC 2 controls). For users without specific data-residency constraints, cloud-first is fine — and it enables features like collaborative review and re-running research from any device — but it’s structurally different from Heptabase’s local-first model and worth naming.

Neither approach is universally correct. Cloud-first wins on collaboration, accessibility across devices, and capabilities like agentic research that need backend compute. Local-first wins on data sovereignty, offline use, and not depending on any single vendor’s continued operation. Pick by what matters more for your specific situation.

Pricing reality check

Heptabase Pro is $11.99/month on annual ($14.99 monthly). Innogath Pro is $9.60/month on annual; Ultra is $32/month on annual. Both have free trials.

The headline ($9.60 vs $11.99) is roughly comparable, but the units of work are different. Heptabase Pro buys long-running knowledge management software with native apps, local-first storage, and unlimited cards. Innogath Pro buys research capacity (5,000 credits/month, ~25 deep research runs) plus the workspace, branching tree, and exports.

For a user who genuinely needs both — research on the discovery side, curation on the long-term side — paying for both is around $21.59/month combined. That’s still cheaper than ChatGPT Plus + most other research tools, and it covers two stages of work that one tool can’t cover well alone.

For a user who only needs one: pick by what your dominant workflow looks like this year. If you’re researching for a thesis or pitching consulting work, Innogath. If you’re maintaining a personal knowledge garden across years of reading, Heptabase. The honest answer for many people is that one of these dominates 80% of their time, and that’s the one to pay for first.

Native app feel vs web app speed

This trade-off is real and underappreciated in tool reviews.

Heptabase ships native apps for Mac, iOS, and Windows. The Mac client in particular has the unmistakable feel of a tool built by people who care about the platform — keyboard shortcuts that follow Mac conventions, Quick Look integration, native fonts, predictable scroll behaviour, working menu bar, sensible window management. For users who spend their working day on Mac, this matters more than feature lists suggest.

Innogath is a desktop web app. It runs in a browser and behaves like a web app: the chrome is browser chrome, scroll behaviour is web scroll, keyboard shortcuts are scoped to the page rather than the OS. The trade-off is that web apps update instantly (no client release cycle), work identically across operating systems, and can integrate with web-based services frictionlessly.

If you reach for native quality as a hard preference, Heptabase wins this comparison cleanly. If you reach for cross-platform consistency and zero-install access from any machine, Innogath’s web-app model wins. Neither is wrong. Worth knowing where you sit on the preference before committing to either.

The honest recommendation

Pick by where you are in the project lifecycle, not by which tool is more popular:

• Discovery and cited deliverables → Innogath. The agent does the research you don’t have time to do manually. The output is a defensible cited report you can hand to a stakeholder. The branching tree maps to how investigations actually grow.
• Curation across years → Heptabase. The whiteboard is your permanent knowledge layer. The cards compound over time. The local-first design respects your data. Innogath is not trying to compete on this surface and would be the wrong tool to use this way.
• Both → realistic for serious knowledge workers. Innogath for the research phase of each project; Heptabase for the long-term layer where you keep the synthesised understanding across projects. The handoff is exporting Innogath’s reports as Markdown and importing into Heptabase as cards.

The mistake is asking either tool to do the other’s job. Heptabase trying to do agentic research is doing it manually, slowly, and without citations. Innogath trying to be a long-term personal knowledge garden is doing it without local-first storage and without the spatial freedom that makes Heptabase work. Match the tool to the stage and both feel correct.