How can we help?

An intellectual fingerprint is an AI-generated analysis of your writing that surfaces your recurring themes, unresolved questions, mental models, intellectual influences, blind spots, and the single core question driving your thinking. It reveals how you think — not just what you know.

Anything you've written works: journal entries, essays, Obsidian vault exports, Notion pages, newsletter issues, Twitter/X threads, blog posts, or talk transcripts. The more you provide, the more accurate the result. A week of notes, 2–3 essays, or 5+ journal entries is a good starting point.

At least 300 words for a basic fingerprint. 600–1,000 words gives a strong result. Over 1,000 words gives an excellent one. Less than 300 words tends to produce shallow or generic output — the AI needs enough signal to find patterns.

You can upload .md (Markdown) and .txt files, or an entire folder of Markdown notes. Wikilinks ([[like this]]) are resolved automatically when you upload multiple files. You can also import directly from Substack by URL, connect Notion via integration token, or paste a URL to any blog post.

You get a public profile page with a shareable link. Rodin immediately shows you your most similar thinkers from the community. You can share your profile link, compare fingerprints with specific people, browse others on the Discover page, and subscribe to a weekly digest of activity on your profile.

Yes. Rodin is currently free to use. You can generate your intellectual fingerprint, create a public profile, and find similar thinkers at no cost.

Rodin's AI reads across your body of work and extracts patterns in how you think — what you return to, what you're genuinely wrestling with, who shaped your thinking, and what your writing avoids. It then runs a proprietary 12-dimensional Cognitive Topology Analysis (CTA) that measures abstract cognitive properties like dialectical style, epistemic approach, and temporal stance.

The quality of the fingerprint depends heavily on the writing you provide. If the result feels generic or off, try: (1) adding more varied writing — mix journals, essays, and notes; (2) using writing where you're actually thinking through ideas, not just recounting events; (3) evolving your fingerprint with fresh writing via "Evolve Your Fingerprint" on your manage page.

Yes. Go to your profile's manage page (use your management link) and click "Evolve Your Fingerprint". This re-runs the analysis on new writing and updates your profile. Your old fingerprint is saved in history — you can see how your thinking has changed over time on your profile page.

CTA is Rodin's proprietary 12-dimensional model that measures abstract properties of how you think — not what you think about. Dimensions include epistemic style (how you form beliefs), temporal stance (whether you think retrospectively or prospectively), reasoning mode (first-principles vs. pattern-matching), and more. These are computed entirely from linguistic patterns in your writing with no external API calls.

The archetype is a 3-word label that captures your dominant intellectual identity based on your CTA profile. It's a shorthand for your cognitive style — useful for quickly recognising patterns across thinkers, not a box to put yourself in.

Accuracy increases with the amount and variety of writing you provide. With a few thousand words of authentic personal writing, most users find the fingerprint remarkably resonant — especially the core question, blind spots, and archetype. More text = more pattern signal = sharper fingerprint.

Matching combines three layers: 512-dimensional semantic embeddings for topical similarity, Jaccard similarity across fingerprint layers (themes, questions, models, DNA) with learned weights, and 12-dimensional CTA vector distance for cognitive style proximity. The final score blends all three.

Your Intellectual Twin is your single closest match — the person whose cognitive fingerprint overlaps most with yours across all layers. The "Why you match" section on their card explains the specific shared ground and productive tensions between your fingerprints.

Each profile shows the person's contact handle (Twitter/X or email) if they've added one. Click "Connect" on their profile to be counted as an interaction, then reach out directly via their handle. You can also use the Compare page to show someone exactly how your fingerprints relate before reaching out.

Discover shows all public Rodin profiles. You can filter by themes, questions, mental models, reasoning mode, or search for specific concepts. It's designed for exploring the community and finding thinkers who work on problems similar to yours.

Yes. On any profile page, click "Compare" to see a side-by-side view of how your fingerprints relate. You can also visit /compare directly and enter two profile IDs. When someone shares a profile link with ?ref=their_id, creating your fingerprint immediately shows the comparison.

Your writing is sent to the AI to generate your fingerprint, then discarded. The fingerprint itself (themes, questions, models, etc.) is stored in your profile. The raw writing text is not permanently retained after generation.

No. Your writing is not used to train AI models. It is sent to our AI system to generate your fingerprint and then discarded. See our Privacy Policy for full details.

Your profile is public and discoverable on the Discover page. Anyone with your profile link can view your fingerprint, name, and contact handle. Your email address (if provided) is never shown publicly — it's only used for management link recovery and the optional weekly digest.

Go to your management page (using your management link) and scroll to "Danger Zone" at the bottom. Click "Delete Profile". This permanently deletes your profile, fingerprint history, interactions, and all associated data. It cannot be undone.

Visit your profile page at rodin.fyi/p/[your-id] and click the settings icon, or go directly to /p/[your-id]/manage. If you're locked out, enter your email address on the unauthorized screen and we'll send you a new management link — as long as you registered with an email address.

Use your management link to access the edit page. You can update your name, contact handle (Twitter/X or email), and the email used for management link recovery.

If you subscribed from your profile page, Rodin sends you a weekly email showing how many people viewed and connected with your profile, plus your closest intellectual match from the community that week.

Click the unsubscribe link at the bottom of any digest email. You can also re-subscribe from your profile page at any time.

Select "Notes" on the home page, then click "Notion". You'll need a Notion Integration Token (create one at notion.so/my-integrations with read access to relevant pages) and the URLs of the pages to import. Rodin fetches the text directly via the Notion API.

Select "Newsletter", paste your Substack URL (e.g. yourname.substack.com), and click "Fetch". Rodin reads your RSS feed and pulls the full text of your most recent posts automatically.

Some sites block automated fetching (paywalls, login-required content, aggressive bot protection). If the fetch fails, copy and paste the text directly into the textarea — it works just as well.

This occasionally happens under load or with very long inputs. Refresh and try again — nothing is lost and the fingerprint generates fresh each time. If it fails repeatedly on the same text, try reducing the amount slightly.

Yes. Upload an entire vault folder of .md files at once — Rodin parses wikilinks and assembles the text automatically. Any markdown-based tool (Logseq, Roam, Bear, Capacities) works the same way.