Skip to content

Hjarni vs Ragie

Ragie indexes uploaded and synced documents for retrieval. Hjarni is a Markdown knowledge base you and your assistants write in directly.

Free to start. No credit card required.

Quick answer

Ragie provides managed indexing and retrieval over files, raw content, and synced sources, with an official hosted MCP Retrieve tool per partition. Its APIs can create and update documents. Hjarni provides Markdown notes you and your assistants edit directly through the app or its official MCP server. Choose Ragie for retrieval over a document corpus and Hjarni for a shared writing surface with folders, history, and AI guidance.

Checked 20 October 2026 against MCP scope and Retrieve tool, Document APIs, Retrieval descriptions, Connectors, Base Chat, Pricing. Plans and features can change.

Swipe the table sideways to compare →

Hjarni and Ragie feature comparison
Feature Hjarni Ragie
Primary surface
Markdown notes in folders Uploaded, API-created, and synced documents
MCP server
Official, hosted Official, hosted per partition
MCP actions
Search, read, create, update Retrieve document chunks
Content updates
Edit the note directly Update APIs and source syncing
Organization
Nested folders and tags Partitions and metadata
Instructions
Folder-level AI guidance Partition retrieval descriptions
Retrieval
Full-text search Hybrid retrieval and reranking
Writing and chat
Notes editor plus connected AI Retrieval APIs and Base Chat
Pricing
Flat plans with note quotas Monthly tier plus metered overages
Portability
Markdown ZIP with active notes and attachments Content/source APIs
Best fit
Documents you maintain in a knowledge base Retrieval over existing or synced data
Details behind the comparison
Primary surface
Ragie supports files, raw text, URLs, and connectors. It is not restricted to manual uploads.
MCP server
Both expose first-party hosted MCP access. Ragie scopes each endpoint to one partition.
MCP actions
Ragie documents a Retrieve tool. Content updates use its APIs or source connectors, rather than that MCP tool.
Content updates
Ragie can update file content, raw text, URL-backed documents, and metadata. Re-uploading is not the only update path.
Organization
Ragie partitions isolate retrieval data. Hjarni folders also provide a document-browsing surface.
Instructions
Ragie can generate a context-aware Retrieve tool description. This differs from instructions for how an assistant should work in a folder.
Retrieval
Choose retrieval behavior using representative questions and documents.
Writing and chat
Ragie also offers a ready-made chat surface; custom application development is not its only workflow.
Pricing
Ragie has a free Developer tier, Starter at $100/month, and Pro at $500/month. Processing, search/storage, and media have distinct allowances or rates.
Portability
Ragie exposes document content and source operations. This is a different exit workflow from Hjarni export.

Hjarni has its own native apps. Free on every plan for iPhone, iPad, Mac, Android and Windows. Same account, same notes as the web. An internet connection is required.

Sources can stay where you maintain them

Ragie accepts uploaded documents, raw text, and URLs, and its connectors sync supported sources. Its API exposes updates for files, raw content, and URL-backed documents. Processing is asynchronous before updated content is ready for retrieval. Hjarni keeps Markdown notes in the knowledge base itself, where people and assistants can update them directly.

MCP retrieval and API writes are different

Each Ragie partition has a hosted MCP server exposing Retrieve, which returns relevant document chunks. Its content-management APIs are a separate surface. Hjarni's MCP tools include creating and updating notes as well as searching and reading them.

Managed retrieval or a writing surface

Ragie handles parsing, indexing, hybrid retrieval, and reranking. Its Base Chat product also provides a ready-made conversational interface, so it is not necessary to build a custom app for every workflow. Hjarni centers on writing and maintaining notes, with full-text search and folder guidance for connected assistants.

Compare the allowances you will use

Ragie's free Developer tier has limits; paid Starter and Pro plans include allowances with additional processing and search/storage charges. Audio and video have separate processing and storage meters. Hjarni uses flat plans with note quotas and no first-party per-query charge. These products cover different work: indexing media and maintaining Markdown notes are not equivalent tasks.

When to use Ragie

  • You want retrieval over an existing document corpus
  • You need source connectors, hybrid search, or multimodal indexing
  • You want retrieval APIs or a chat interface over that corpus

When to use Hjarni

  • You want to write and edit Markdown notes directly
  • You want assistants to create and update those notes through MCP
  • You want folders, note history, and folder-level AI guidance

Ragie indexes your source material. Hjarni is a shared place to maintain notes.

Give your AI a memory

Common questions

Common questions

What is Ragie?

Managed retrieval infrastructure with ingestion, connectors and document-management APIs. Its native partition MCP exposes retrieval, which is different from a note-editing MCP tool set. See official documentation.

Can Ragie documents be updated?

Yes. Supported APIs and connectors can update Ragie documents. Its native MCP endpoint handles retrieval, and Base Chat provides a conversational interface.

Which is cheaper?

Compare the actual workload and product tier. Ragie has separate ingestion, storage or retrieval-related meters depending on its offering. Hjarni uses note-based hosted plans. The cost comparison depends on the amount and type of content you process.

Can I migrate from Ragie to Hjarni?

Check how to retrieve original documents and metadata, then convert compatible content into notes. A chunk index, citations and connector metadata do not automatically become an identical Markdown knowledge base.

Start here

Write once. You both remember.

25 notes free, forever. No credit card required.

5 out of 5 from 14 reviews 14,500+ notes a day read or updated by AI

Works with Claude, ChatGPT, Cursor and any MCP client.