Skip to content

FlareMo

Memos-style quick notes and a shared team knowledge base, with semantic search and long-term memory for AI agents

FlareMo timeline on desktop and phone with the note composer, a note with a highlighted quote, its tags, and counters for memos, tags and days

FlareMo brings the quick-capture note timeline of Memos and flomo to a single Cloudflare Worker, with an optional team mode. Search combines SQLite full-text and Vectorize lookup. A Memos-compatible API, a CLI and MCP endpoints let scripts, Memos clients and AI agents read and write the same notes.

Notes in FlareMo carry tags, Markdown, image and audio attachments, and a revision history you can restore from. Search runs on SQLite FTS5 with operators such as has:attachment and before:YYYY-MM-DD. Daily Review, Random Walk and related-note suggestions bring older entries back up. Projects add a kanban board with priorities and due dates, and overdue tasks raise in-app notifications, with optional Web Push. Single-user mode is the default. Team mode adds owner, admin and member roles, one-time invite links, time-boxed reader seats and three visibility levels: private, team and public share link. Better Auth handles username and password login. Revocable memos_pat_ tokens let scripts and Memos clients such as Moe Memos call the /api/v1 endpoints. You can import from Memos or flomo, export everything, install the PWA to queue edits offline, and extend note cards with sandboxed plugins. AI coding agents can use FlareMo as long-term memory. Through the flaremo CLI, a flaremo-memory skill and MCP endpoints, Claude Code, Cursor or Codex record preferences and project decisions, which you can review, correct or lock at /memory. The Worker serves the React frontend as static assets. It binds D1 for notes and settings, R2 for attachments and export bundles, two Queues for member removal and data exports, two Vectorize indexes fed by Workers AI embeddings, a rate limiter and a daily Cron Trigger. Besides the Deploy to Cloudflare button, a GitHub Actions workflow can deploy from a fork.

Built for: Individuals who keep short notes in Memos or flomo, small teams that want a shared knowledge base, and developers who want persistent memory for AI agents.

What you can use it for

  • Capture short personal notes with tags, search and attachments
  • Run a team knowledge base with private, team and public visibility
  • Give AI coding agents long-term memory through a CLI or MCP
  • Self-host a backend for Memos mobile clients such as Moe Memos
  • Move notes out of Memos or flomo

Deploy

  • One-click · main path

    Branch main

  • CLI

    Branch main

    Set FLAREMO_PUBLIC_URL in wrangler.jsonc before deploying. After the deploy, open /setup and enter FLAREMO_BOOTSTRAP_SECRET to create the owner account.

    pnpm install
    cp wrangler.jsonc.example wrangler.jsonc
    pnpm provision:remote
    pnpm exec wrangler secret put BETTER_AUTH_SECRET --config ./wrangler.jsonc
    pnpm exec wrangler secret put FLAREMO_BOOTSTRAP_SECRET --config ./wrangler.jsonc
    pnpm deploy:dry-run
    pnpm run deploy

The button provisions all resources. Set FLAREMO_PUBLIC_URL and BETTER_AUTH_SECRET, redeploy, and create the owner at /setup.

What it needs to run

  • Workers AIAI
  • D1 databasesflaremo
  • R2 bucketsflaremo-attachments
  • Vectorizeflaremo-memories, flaremo-memos

Good to know

The deploy docs describe the Deploy to Cloudflare button as community-supported and best for quick trials, with manual deployment through Wrangler as the fully supported path. First attempts with the button can fail with a GitHub API rate limit error and succeed after a short wait. The Memos-compatible API implements a subset of Memos Server, and the docs do not promise full parity with it or with third-party clients. Under AGPL-3.0, anyone who runs a modified version as a network service must offer its source code to that service's users.

Cost on Cloudflare

Likely fits the free plan

The README targets a free Cloudflare account, and the default FLAREMO_VECTORIZE limits match the Vectorize free quota: 5 million stored vector dimensions and 30 million queried dimensions per month. With the default 1,024-dimension embeddings, the stored allowance holds about 4,900 vectors. Workers Free also includes D1 (5 GB, 5 million rows read and 100,000 written per day), Queues (10,000 operations per day) and 10,000 Workers AI Neurons per day.

Source

The estimate is based on Cloudflare’s documented limits and the app’s configuration. What you pay depends on your usage and plan.

Where this comes from

An alternative to Evernote, flomo, Google Keep, Mem, Mem0, Simplenote.

Listing history

  • Oct 2, 2026 · FlareMo listed
  • Inkstone public demo sign-in screen with the demo account filled in

    Private Markdown notebook with backlinks, offline sync, MCP access and WebDAV or S3 backups, running on Workers

    Deploy
    Manual
    License
    LGPL-3.0-only
    GitHub stars
    1.1k stars
    Last commit
  • Obsidian note on desktop with two live cursors and the YAOS CRDT: Connected status in the footer

    Live Obsidian vault sync over Yjs CRDTs, served by a Worker you deploy and claim in your own Cloudflare account

    Deploy
    One-click
    License
    0BSD
    GitHub stars
    1k stars
    Last commit
  • Second Brain dashboard Projects tab with a new project form and five projects with memory counts

    Self-hosted memory store that Claude, ChatGPT, Cursor and other MCP clients share, with recall by meaning

    Deploy
    One-click
    License
    MIT
    GitHub stars
    802 stars
    Last commit