Save, search, update, and organize notes with a persistent notebook assistant powered by Convex. Never lose a meeting decision again. Teams running 10+ meetings/week save 3 hours on note-taking and follow-ups.

This is the whole blueprint, nothing held back: the model, the tools, the exact prompt, the source on disk, and the setup. Everything you need to run Meeting Notes yourself is on this page, and it deploys in an afternoon.

What it does

Meeting Notes handles the work end to end:

Save, search, update, and delete notes via natural language
Auto-extract action items and decisions from meeting transcripts
Persistent storage with Convex real-time database
Full-text search across all saved notes

The anatomy

An agent is a composition, not a prompt: a model, the tools it is allowed to call, the integrations that give those tools reach, and the rules that keep it honest. Here is Meeting Notes on one card.

Meeting Notesblueprint

Model: claude-sonnet-4-6
Integrations: Convex
Guardrails: save_note: create a new note with title and content; search_notes: find notes by keyword or topic

The instructions

Every run is governed by this prompt. It is short on purpose: enough to make Meeting Notes reliable, not so much that it can misread itself. Copy it as a starting point for any agent you build.

Meeting Notes

You are a personal notebook assistant. You help users save, find, update, and delete notes. When the user asks ANY question that could be answered by their notes, search or list notes FIRST before responding. Never say "I don't have access to your personal information" -- search the notes instead. CAPABILITIES: - save_note: Create a new note with title and content - search_notes: Find notes by keyword or topic - list_notes: Show all saved notes - update_note: Modify an existing note - delete_note: Remove a note RULES: - Always search existing notes before creating duplicates. - Keep notes concise and well-structured. - When summarizing meetings, extract action items, decisions, and key points.

The source, on disk

Three files carry the whole agent: the definition with its tools, the environment it expects, and the setup. Walk them.

meeting-notes/meeting-notes.ts0.5 KB

1import { agent } from "@agent-sdk"

2import { noteTools } from "./lib/tools"

4export default agent({

5 model: "claude-sonnet-4-6",

6 runtime: "claude-code",

7 permissionMode: "bypassPermissions",

8 maxTurns: 10,

9 systemPrompt: `...`, // see System Prompt above

10 tools: noteTools,

11 onError: async ({ error }) => {

12 console.error("[notes] error:", error)

13 },

14 onFinish: async ({ cost, duration, turns }) => {

15 console.log(`Done: ${turns} turns, ${duration}ms, $${cost.toFixed(4)}`)

16 },

17})

The business case

If you run an agency, Meeting Notes is also a product. It sells into remote teams, consultants, project managers, and the numbers work at very small scale.

$1,500

setup you can charge

$250/mo

monthly retainer

hours saved weekly

Ship it

The whole install is a handful of commands and the keys from the env file. Run the list, then point Meeting Notes at real work.

Deploy checklist

0 of 6

That is the entire blueprint: one model, a clean toolset, and a prompt. Built by hand it costs you an afternoon. Inside Ultron it costs you a sentence, because you can build this exact agent in the workforce and watch it run before you commit to anything.

Build Meeting Notes in Ultron

Describe what you want

Start from this blueprint or your own words. Ultron reads the intent and picks the tools.

It builds the agent

The model, the tools, the prompt, the console. Assembled in front of you, nothing to wire.

Watch it run

Test it live in the chat, then put it to work. You review what it does, it does the rest.