Hjarni
Hjarni is a personal knowledge management platform for organizing, searching, linking, and maintaining notes, containers, and tags.
Verdict
Common use cases
- Save meeting notes automatically after standups
- Archive research findings by topic container
- Tag customer feedback for later review
- Capture brainstorm ideas during design sprints
- Retrieve past decisions when planning features
Integration
- Vendor
- Hjarni
- Category
- other
- Auth
- API_KEY
- Tools
- 15
- Composio slug
hjarni
Tools
- Create Container
Create a root or nested Hjarni container for organizing notes.
- Create Note
Create a Markdown note in the inbox or a container. Tags named in tag_names are applied as part of creation. Free accounts are limited to 25 notes, so check account status if quota pressure is possible.
- Create Notes
Create 1-25 Markdown notes in order and return one result per input. This batch is non-atomic: successful notes remain created if another item fails.
- Create Tag
Create a Hjarni tag for organizing notes.
- Get Account Status
Return the connected Hjarni identity, token scope, plan, teams, and current resource usage and quotas. Use this before creating many notes or attachments; Free accounts allow 25 notes, 5 attachments, and 20 MB of attachment storage.
- Get Note
Get one Hjarni note by ID with its Markdown body, tags, container, attachments, links, and optimistic-lock version.
- List Containers
List Hjarni containers as active roots, archived roots, or all active containers in a flat view. Use next_cursor to page through results.
- List Notes
List and filter personal Hjarni notes, including active, inbox, archived, or favorited notes. Use next_cursor to page through results; use query for simple personal-note filtering.
- List Tags
List all Hjarni tags and their note counts. Use next_cursor to page through results.
- List Trashed Notes
List recoverable trashed Hjarni notes with their deletion and scheduled purge times. Use next_cursor to page through results.
- Move Note to Trash
Move a note to recoverable Trash. This is not immediate permanent deletion: Hjarni retains the note for about 30 days, and RESTORE_NOTE can recover it before purge.
- Rename Tag
Rename an existing Hjarni tag without changing its note associations.
- Restore Note
Restore a note from Trash before its scheduled purge and return the active note.
- Update Container
Update a Hjarni container's name, description, parent, position, or container-specific AI instructions.
- Update Note
Update note metadata, placement, tags, or body. body replaces the entire Markdown body; append_body safely appends text. Provide expected_lock_version from GET_NOTE to prevent overwriting concurrent edits.
Setup
Setup guide
- 11. In Switchy, open your workspace settings and click 'Add Integration' under the MCP section. 2. Select Hjarni from the catalog and choose 'Connect with API Key'. 3. Open Hjarni in a separate tab, navigate to Settings → API Keys, and generate a new key with read-write scope. 4. Paste the key into Switchy's prompt and click 'Authorize'. 5. Switchy will confirm the connection and display your Hjarni account plan and current note count. 6. Inside any Project, type '@Hjarni check my account status' to verify the integration works. 7. To create your first note, try '@Hjarni create a note titled "Project kickoff" with the body: We agreed on three milestones today.' 8. The AI will return the note ID and confirm it landed in your inbox; open Hjarni to see it appear.
What teammates see: by default, memories from Hjarni are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Check Note Quota
@Hjarni get my account status and tell me how many notes I have left on my current plan.
Save Meeting Summary
@Hjarni create a note titled 'Sprint Planning – [today's date]' with a summary of this conversation, and tag it 'meetings' and 'sprint-12'.
Organize Research by Topic
@Hjarni create a container called 'Competitor Analysis' and then create a note inside it titled 'Feature comparison' with a table of the features we discussed.
Batch-Create Action Items
@Hjarni create five notes from these action items: 1) Draft onboarding email, 2) Schedule user interviews, 3) Update pricing page, 4) Review analytics dashboard, 5) Prep demo script. Tag them all 'Q1-launch'.
Retrieve Past Notes
@Hjarni list all notes tagged 'customer-feedback' from the last month and summarize the top three themes.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Hjarni list all notes tagged 'meeting' from the last week and show me their containers
Found 4 notes tagged 'meeting' from the past 7 days: 1. "Q1 Planning Session" (Container: Projects/2024-Q1) – created Jan 8, 847 words 2. "Design Review Notes" (Container: Product/Design) – created Jan 10, 412 words 3. "Sprint Retro" (Inbox) – created Jan 11, 203 words 4. "Client Sync - Acme Corp" (Container: Clients/Acme) – created Jan 12, 591 words All notes are active. The Sprint Retro is still in your inbox and may need filing into a container.
This example shows Hjarni's list-and-filter capability across tags, dates, and container hierarchy. The MCP returns metadata (word count, container path, creation date) but not full Markdown bodies unless you request a specific note by ID. Free accounts are capped at 25 notes total, so filtering becomes critical as you approach quota.
@Hjarni create a note titled 'API Integration Checklist' in the Engineering container with tags 'backend' and 'onboarding'
Created note "API Integration Checklist" (ID: nt_8x3k9m2p) in container Engineering. Applied tags: backend, onboarding Status: Active Word count: 0 (empty body) Created: Jan 15, 2025 at 10:42 AM UTC The note is now available in your Engineering container. You can add Markdown content by updating the note or appending via a follow-up prompt.
This demonstrates Hjarni's write capability: creating a note with metadata (title, container, tags) in one call. Tags are applied atomically during creation. Free accounts hit a 25-note hard limit, so check account status first if you're batch-creating notes. The MCP confirms the note ID for subsequent updates or linking.
@Hjarni check my account status, then draft a summary of all notes in my 'Research' container and create a new note called 'Research Digest' with that summary
Account: Free plan (18 of 25 notes used, 2 of 5 attachments used). You have 7 notes remaining before quota. Found 5 notes in Research container: - "LLM Context Windows" (1,203 words, tagged 'ai') - "Vector DB Comparison" (847 words, tagged 'infrastructure') - "Prompt Engineering Patterns" (612 words, tagged 'ai') - "Cost Analysis - Embedding APIs" (391 words, tagged 'infrastructure', 'cost') - "RAG Architecture Notes" (1,104 words, tagged 'ai', 'architecture') Created note "Research Digest" (ID: nt_7k2m5n8q) in Research container with a 340-word synthesis covering vector database trade-offs, embedding cost benchmarks, and RAG design patterns extracted from the five source notes. Tagged 'summary'.
This example chains three operations: quota check, multi-note read, and AI-assisted synthesis into a new note. It shows how Hjarni's account-status tool prevents quota violations and how the AI can reason over note metadata and bodies to produce derivative artifacts. The batch-create tool supports up to 25 notes at once, but this single-note creation is safer near quota limits.
Use-case deep-dives
When Hjarni beats shared docs for remote standups
A 6-person engineering team across three time zones runs async standups in Slack but loses context in thread sprawl. Hjarni's container hierarchy and tagging let each engineer drop standup notes into a shared "Weekly Standups" container, tagged by sprint and blocker status. The batch-create tool writes 5 notes at once when someone dumps their week's updates. The quota matters: free accounts cap at 25 notes, so a team burning through daily standups hits the wall in a week. If your team writes fewer than 20 structured updates per month and needs searchable history without Notion's page-load tax, Hjarni's API-first design is the call. Upgrade to paid if you're logging more.
Hjarni as a lightweight ticket-to-insight pipeline
A 3-person support team at a SaaS startup collects feature requests in Intercom but has no budget for a dedicated insight tool. They pipe high-signal customer quotes into Hjarni notes, tagging by product area and urgency. The list-notes filter pulls all "urgent + billing" tags for the weekly product sync. The 5-attachment limit on free accounts means you can't dump screenshots for every ticket, but text snippets and links work fine. This setup breaks down when you're triaging more than 100 feedback items per month or need rich media; at that scale, a proper CRM integration wins. For early-stage teams capturing qualitative signal without Airtable overhead, Hjarni's markdown-native structure keeps it fast.
When Hjarni replaces the retro Google Doc
A 5-person product squad runs biweekly retros and historically loses action items in a shared Google Doc that no one revisits. They switch to Hjarni: one container per sprint, one note per action item, tagged by owner and theme. The get-account-status tool checks quota before batch-creating 12 action items from the retro. Two weeks later, list-notes filtered by "open + sprint-23" surfaces what's still pending. The model falls apart if your retros generate 30+ action items per cycle or you need Gantt-chart dependencies; Linear or Asana own that complexity. For teams that want structured retro follow-through without project-management ceremony, Hjarni's tag-and-container model is the lightest viable path.
Frequently asked
What does the Hjarni MCP do in Switchy?
It lets AI agents create, read, and organize Markdown notes in your Hjarni account. Agents can build containers (folders), apply tags, batch-create up to 25 notes at once, and retrieve notes by ID or filter. Think of it as giving your AI direct write access to your personal note system, so it can capture research, draft ideas, or log decisions without you copy-pasting.
Do I need a paid Hjarni plan to use this MCP?
No, but free Hjarni accounts cap you at 25 notes and 5 attachments total. The MCP includes a Get Account Status tool so agents can check your quota before trying to create notes. If you hit the limit mid-workflow, the agent will see the error and can warn you. Paid plans remove these caps, making the integration more useful for heavy note-taking.
Can the MCP search across all my existing Hjarni notes?
Partially. It can list notes with basic filters—inbox, archived, favorited—and supports a simple query parameter for personal notes. It cannot run full-text search across note bodies or perform complex boolean queries. If you need deep search, use Hjarni's web app; the MCP is better suited for creating and retrieving notes the agent already knows about by ID or tag.
How is this different from just asking AI to write notes and pasting them into Hjarni myself?
Speed and context. The agent writes directly into your Hjarni containers, applies tags, and can batch-create 25 notes in one call—no clipboard juggling. It also reads your existing notes, so it can reference past entries, update tags, or organize new notes into the right container automatically. Manual pasting loses that closed-loop workflow.
Who on my team should connect the Hjarni MCP?
Whoever owns the Hjarni account you want agents to write into. Each Switchy user connects their own Hjarni API key, so notes land in that person's workspace. If your team shares a Hjarni account, one person connects it; if everyone has individual accounts, each connects their own. The MCP doesn't sync notes between team members—it's single-user per connection.