Cairni — Complete Overview

At its heart, Cairni answers a practical problem: most people already have the knowledge they need, scattered across PDFs, notes, recordings, and browser tabs — they just can't access it quickly or reliably. Cairni's job is to gather all of that material in one place and turn it into something genuinely useful: a connected, searchable, citable wiki that an entire team can share.

What sets Cairni apart from a general-purpose AI chat tool like ChatGPT is persistence and sourcing. A general chat assistant has no memory of your documents between sessions and produces answers from its pre-trained knowledge with no traceable citation. Cairni, by contrast, writes actual wiki pages grounded in *your* material, cross-links concepts and entities automatically, and ensures that every fact can be traced back to the source it came from. The FAQ goes deeper on this distinction, including how Cairni differs from RAG (Retrieval-Augmented Generation) — which retrieves document chunks at query time but produces no lasting, navigable artifact.

Everything in Cairni flows through a three-step loop described in detail on How Cairni works.

Step 1 — Add sources. You bring in the material you want Cairni to learn from. This can be uploaded files, pasted text, web URLs, connected Notion pages, or recorded audio. The full breakdown of supported formats lives on What you can add.

Step 2 — Compile. Once a source is saved, Cairni reads it and automatically summarises it, writes or updates concept and entity pages, cross-links related ideas, and maintains an index so every claim traces back to the original. You never have to write a single page yourself — though you can always edit what the AI produces. This compilation step is what distinguishes Cairni from a simple document store: the output is a structured, navigable knowledge base, not a pile of raw files.

Step 3 — Ask. Open the wiki and type a question in plain language. Answers are grounded in your compiled pages and carry citations you can click through to verify. Crucially, if a generated answer is particularly good, you can save it back as a new wiki page — meaning the wiki grows smarter the more you interact with it. The loop then continues: new pages become part of the knowledge base that future answers draw on.

Cairni accepts a wide variety of source types, making it practical to use across different workflows. As covered on What you can add, you can upload files in PDF, Markdown, plain text, or Word format; paste text directly; import a web URL; connect a Notion workspace; or add audio recordings such as meetings or voice notes (Cairni transcribes them first, then compiles the transcript like any other source). You can freely mix source types within a single notebook — for example, combining a set of academic PDFs with a recorded team discussion and a pasted article from the web. [Cairni guide.md]

Key features covers each of these in depth, but here is the full picture in one place:

• Cited answers. Every answer — whether from navigating the wiki or from the chat assistant — links back to the exact source page it came from. No black-box claims.
• Assistant & Memory. A daily chat assistant answers questions grounded in your wikis. Alongside it, Cairni maintains a private Memory wiki — a page about you that it builds as it learns your context, and which you can read and edit directly.
• Teams. You always start as a team of one. When you're ready, you can invite colleagues into a shared workspace with per-notebook access control, approvals, and comments, keeping knowledge changes deliberate and trackable.
• Public wikis. Any wiki can be published publicly with a single click, making it readable by anyone — including search engines and AI answer engines. Useful for open research, documentation, or community knowledge bases.
• Export / own your data. You can export your wikis at any time. Your knowledge stays yours, regardless of which plan you are on.
• MCP. Cairni exposes a Model Context Protocol interface so that AI agents and external tools can connect to and query your wikis programmatically. [Cairni guide.md]

Use cases walks through the four main patterns in detail:

Research — Drop in papers, articles, and notes on a topic. Instead of hunting through PDFs, ask questions and receive cited answers drawn from your compiled sources.

Learning — Turn lecture slides, reading notes, and course materials into a study wiki. Every page links back to its source, so you can trace any claim to the original text and build on your understanding iteratively.

Meetings & decisions — Add recordings, transcripts, or written notes from meetings. Cairni compiles them into a searchable record so you can find what was decided, who said what, and why — without digging through old files.

Team knowledge — Build a shared, always-current wiki instead of scattered docs and chat threads. Invite teammates, control access per notebook, and let everyone ask questions grounded in the same sources. [Cairni guide.md]

The full walkthrough is on Getting started, but the path is straightforward:

1. 1.Create a notebook — a self-contained wiki organized around one topic or project.
2. 2.Add your sources — upload files, paste text, import URLs, connect Notion, or record audio.
3. 3.Ask the wiki a question — start broad (*"Give me an overview of everything here"*) or specific; save useful answers back as new pages.
4. 4.Invite teammates or publish — share access with colleagues, or publish the wiki publicly with one click.

Because every source is traced and every answer is cited, you can hand a notebook to a colleague or make it public with confidence that the content is accountable. [Cairni guide.md]

Pricing covers the details, but here is the shape of the model. Cairni is free to start. Paid plans are credit-based — credits are consumed when Cairni compiles sources and generates answers. [Cairni guide.md]

The Free tier lets you try Cairni without commitment. Pro ($20/month), Plus ($40/month), and Max ($100/month) serve individual users with increasing credit volumes. The Team plan ($30 per seat/month) adds the shared workspace features — per-notebook access control, approvals, and comments — for groups. Check Pricing for the most current credit allocations, as those details can change. On every plan, data export is available and your knowledge stays yours. [Cairni guide.md]

The FAQ addresses the most frequent questions new users bring, including:

• How is Cairni different from ChatGPT? — ChatGPT is a general-purpose chat with no persistent memory of your documents. Cairni builds a growing, citable knowledge base from your own sources that you can read, edit, and share.
• How is this different from RAG? — RAG retrieves document chunks at query time and returns a chat answer with no lasting artifact. Cairni writes actual wiki pages, cross-links them, and lets you save good answers back — the output is a navigable wiki, not just a response.
• Who writes the wiki pages? — Cairni's AI writes them automatically when you add sources. You can edit the results, but you don't have to.
• Is my data mine? — Yes. You can export your wiki at any time.
• Can a team share one wiki? — Yes. The Team plan supports shared workspaces with per-notebook access control, approvals, and comments. [Cairni guide.md]

This page gives you the full picture. Every section links to a dedicated page where you can go deeper:

• Welcome to Cairni — a warm, quick introduction and orientation.
• How Cairni works — the ingest → compile → ask loop in detail.
• Getting started — step-by-step from blank notebook to living wiki.
• What you can add — every supported source format.
• Key features — cited answers, Assistant & Memory, teams, public wikis, export, and MCP.
• Use cases — research, learning, meetings, and team knowledge.
• Pricing — Free, Pro, Plus, Max, and Team plans with credit-based billing.
• FAQ — comparisons, data ownership, team sharing, and more.

Or simply type your question directly into the chat — the wiki will answer you with citations.