nocal.app/mcp

The notes app for AI agents, built on MCP.

nocal turns notes into an active memory layer your tools can read and write. Ask Cursor, ChatGPT, or Claude to do work, then save every decision back into structured notes automatically.

Download nocal Connect Agents

How MCP flows

nocal is the writeback layer in your MCP stack.

Step 01

Agent context retrieval

Cursor, Claude, or ChatGPT calls nocal MCP tools to find the right note history and current project context.

Step 02

Reasoning and synthesis

Your model turns fragmented context into clear outputs: release notes, plans, summaries, and action lists.

Step 03

Structured writeback

The output is written back into nocal notes with safe patch semantics, so work survives beyond chat history.

Use cases

Real workflows, built on your notes.

Cursor

Update release notes on every commit

Ask Cursor to read your branch and update your release note — no manual writeup required.

  • Reads commits, diffs, and related notes
  • Generates a clean user-facing summary
  • Appends directly to your shipping note
Ship prep: release notes + changelog sync
+...×
summarize this branch and update my v3.1 release note in nocal

Planning tool calls 4s

I'll find your existing release note, extract user-facing changes from this branch, then append release summary + QA checklist into the same note.

Explored 7 files, 3 searches

↳Ran Notes Search in nocal
↳Ran Notes Get + Notes Patch in nocal

Result: v3.1 shipping note updated

  • Read: notes_search + notes_get on "v3.1 release prep"
  • Write: notes_patch appended release summary and QA checklist.
  • Output: one canonical note ready for changelog + team review.
How can I help you today?
How can I help you today?
Sonnet 4.6
This Week
Deploy Guide.md
Zero-Downtime Deployment

Claude

Research a topic, save the output

Ask Claude to research anything and write the findings directly into a structured nocal note — checklist, summary, or action items included.

  • Claude reads and writes via MCP
  • Findings saved as structured notes
  • Ready to reference in any future session

Context that carries across sessions

Start research in ChatGPT, write findings to a note. Resume in Claude the next day — same context, no re-explaining.

Every agent on the same page

Cursor writes the branch summary. Claude reads it to draft the spec. One shared note keeps the whole team aligned.

AGENTS.md

Add nocal to your agent rules.

Drop these into your AGENTS.md or .cursorrules to make nocal part of every session.

Release notes
After completing any task, search nocal for a release note
matching the current semantic version (e.g. "v2.4.0 release prep").
Append a brief summary of user-facing changes and any migration notes.
If no matching note exists, create one.
Roadmap & next steps
Before starting a new task, search nocal for a note titled
"roadmap" or "backlog". Read open items and prefer work that
unblocks the highest-priority thread. After completing work,
update the note to reflect what shipped and what’s next.
Decision log
When making an architectural or product decision, find or create
a note named "<project> decisions". Append the decision made,
the alternatives considered, and the reason for the choice.

Prompt ideas

Copy, paste, run.

One-off prompts for Claude or ChatGPT to pull from or write back to your notes.

Claude: summarize open threads
Search my notes for anything marked "open", "TBD", or "unresolved".
Group by project and summarize:
- what’s blocked and why
- what needs a decision
- what can move forward now
Write the summary into a note called "Weekly open items".
ChatGPT: build a task list from context
Read my Q1 retro notes and current roadmap note.
Based on what shipped and what didn’t, generate a prioritized
task list for next sprint. Save it as a new note called
"Sprint [N] tasks" with priority, owner placeholder, and scope.
Claude: research a topic and save findings
Research best practices for [topic].
Read any existing notes I have on this subject first.
Then write a structured note with:
- key findings and recommendations
- tradeoffs or open questions
- 3-5 concrete next steps
Save it as "[topic] research" in nocal.

Future-ready notes

AI can reason anywhere. nocal makes that work reusable.

Capture what your agent did, what it decided, and what ships next. nocal keeps your MCP workflows organized, searchable, and ready for the next run.

Download nocal Connect Agents