Live demo · sign in as admin, break anything — DB wipes every 4h admin access · resets 4h
next reset in 02:14:37 Launch demo
Docs
v1.4 · shipped apr 2026

Ship your agentic startup
by the weekend.

From hackathon prototype to production SaaS — RED gives you the Agentic-Native stack in 90 minutes. So your next 3 months build the product, not the plumbing.

what's in the kit
ai module
Agent runtime
Persistent streaming, tool calls, queues, boundaries, sync + async tasks.
tenancy
Multi-tenant + RBAC
Orgs, roles, invitations, scoped data. Multi-tenant from line one.
billing
Metered billing
Autumn + Stripe. Plans, entitlements, per-agent-run usage tracking.
auth
Better Auth
Email, passkeys, 2FA, OAuth, org invites, role-based access.
admin
Admin dashboard
User management, org oversight, audit logs, billing sync, feature flags.
security
Security by default
Rate limiting, API key scoping, audit trails, durable workflows.
email
Transactional email
Resend + React Email. Invitations, password resets, notifications.
launch
Waitlist
Built-in flow with admin approval. Collect early users, stage access.
/ the stack — serious tools, sensible defaults
full list ↓
built on · convex
The reactive backend that makes the AI Module possible.

Server functions, ACID transactions, CQRS, realtime subscriptions. Every feature of the AI Module is a thin, type-safe layer on top.

Convex· React 19· Bun· Vite· Tailwind v4· shadcn/ui· Better Auth· Zod· Autumn + Stripe· Resend· Lingui i18n· Biome· React Router 7· TypeScript strict· Convex· React 19· Bun· Vite· Tailwind v4· shadcn/ui· Better Auth· Zod· Autumn + Stripe· Resend· Lingui i18n· Biome·
/ 01 · the ai module

The engine every
LLM wrapper wishes
it had.

Most "AI starter kits" give you a streaming chat input and call it a day. RED ships five production-grade primitives built on Convex's reactive database.

red · agent workspace
Atlas · Due-diligence boundary: Atlas Acquisition
websocket · live
Run the full financial deep-dive on Atlas. Take your time.
agent
task · running
eta 00:08:41
deep_research · Atlas Corp financials, 10-K + 10-Q, competitive landscape
external · unlimited runtime 12%
marie · typing In parallel, pull the cap table
queued · 2 | Messages queue while the agent streams — never dropped.
/ 01.a · primitives
5 concepts
01 · Streaming ai-module

Multi-party live streaming over websockets.

Everyone in a thread sees the same tokens, the same typing indicators, the same task updates — in real time. Built on Convex's reactive queries, not a bolted-on socket server.

ws://thread/atlas-14
pierre: 'what about Q3...
agent: 'Q3 revenue was $
marie is typing…
02 · Queue ai-module

Prompt queue that never drops a message.

Users type while the agent streams. Messages land in an ordered queue, processed with back-pressure. No lost turns, no race conditions, no rebuilds.

msg_412processing
msg_413queued
msg_414queued
03 · Boundaries ai-module

Group threads into projects with their own brain.

A Boundary is a first-class container: dedicated memory, scoped knowledge base, members, permissions.

Atlas Acquisition
312 docs · 4 members
Internal R&D
58 docs · 2 members
04 · Agents ai-module

One abstraction for every agent you'll ship.

Tools, memory scopes, models, guardrails, fallbacks — declared, not wired. Spin up a new agent in a file.

agent("researcher", {
  model: "gpt-4.1",
  tools: [search, scrape],
  memory: "boundary",
  tasks: [deep_research],
})
05 · Sync Tasks ai-module

Under 5 minutes. Inline. Blocking.

Quick tool calls that finish inside the turn. Standard chat affordances. Nothing exotic.

fetch_weather · 420ms
query_db · 1.2s
search_web · 3.8s
/ 01.b · the headliner
the headliner · external async tasks

Long-running
without locking

Sync tasks finish inside a turn. External tasks run for hours, off the main thread, and push progress back in real time. Your user closes the tab. Your agent keeps going.

Deep research. Batch scraping. Multi-step agent pipelines. Video renders. Fine-tunes. Things LLM wrappers can't do.

unlimited runtime push on complete progress streaming retry + resume structured results
task.lifecycle
Agent spawns external task
deep_research · Atlas Corp
Runs off-thread, unlimited runtime
no timeouts · no 'are you still there?'
Streams progress back via websocket
34% · 67% · 92%
Notifies the main thread on completion
push · email · in-app
Delivers structured result
report.md · citations · artifacts
most valuable feature
without RED
  • Roll your own websocket layer. Hope it scales.
  • Bolt a queue onto Redis. Debug race conditions at 2am.
  • Long task? Workers, schedulers, cron, a new ops surface.
  • Multi-user threads? Good luck with memory scoping.
  • Three weeks before you write your first prompt.
with RED
  • Reactive queries stream to every client. Zero config.
  • Queue is a Convex table. Ordered, durable, observable.
  • task("async", …) — runs forever, pings on done.
  • Boundaries give every project its own brain.
  • First prompt in production: 90 minutes after bun install.
/ 02 · the philosophy
Every SaaS is becoming agentic. Start there.

Not a boilerplate.
A singularity.

Most starter kits try to please everyone — five databases, three ORMs, four CSS systems, ten abstraction layers. You spend the first week deleting what you don't need, and the second week realizing you can't.

RED picks a side. One stack, chosen on merit. Open the codebase and change anything — nothing is hiding.

principle 01

Highly opinionated, highly personal.

One runtime. One database. One router. One component library. Picked because they're the best at their job today — not because we wanted to sell optionality.

principle 02

AI-native. By design.

RED isn't a SaaS kit with AI glued on. The AI Module is the spine — auth, billing, i18n, routing all wrap around it.

/ 02.a · the RED cli
no clone · no fork · no copy-paste

One command.
Running app.

Buy the license, drop in the key, run the CLI. No GitHub invites, no repo to clone, no setup wiki.

license-gated always latest < 60 seconds
~/code — zsh
$ bun create red-app my-saas \
    --license YOUR_LICENSE_KEY

 verifying license ………………………… ok
 pulling RED@v1.4
 scaffolding ./my-saas
 installing deps with bun
 provisioning convex dev deployment
 seeding boundaries + demo agents
 running on http://localhost:3000

$ cd my-saas && bun red update
 v1.4 → v1.5 · 6 conflicts auto-resolved
same cli · bootstrap + upgrade
bun create red-app
bootstrap

cli setup no git clone.

Validates your license, pulls the latest tagged release, scaffolds the project, wires Convex, and starts the dev server.

unique to RED
bun red update
migrate

Update with Skill.

Reads the diff between your release and the new one, applies it to your codebase, and resolves conflicts with your own edits.

+ ai coding skills
red-init  ·  red-migration
Claude / Cursor skills that wrap the CLI — so your agent can install and upgrade RED on your behalf, in natural language.
skill docs →

built for the agent workflow

/ 03 · deploy

Two ways.
Both free to start.

Ship to a generous cloud free tier today. Or self-host on your own box. Same codebase, either way.

recommended
option a · cloud

Convex + Cloudflare.
Free tier, forever-ish.

Backend on Convex (reactive DB, server functions, cron, file storage). Frontend on Cloudflare Pages (global edge, free SSL, preview URLs).

backend
Convex
1M fn calls · free
frontend
Cloudflare
unlimited bandwidth · free
git push → deploy
preview URL per PR
scale up by upgrading, not rewriting
open-source path
option b · on-premise

Self-host.
Your hardware. Your rules.

Convex is open-source, so is everything else in the stack. Docker-compose onto a VPS, an air-gapped server, or bare metal.

$ docker compose up -d
 convex backend      :3210
 app (vite/bun)      :3000
 postgres            :5432
 RED running on-prem
no vendor lock-in
SOC2 / HIPAA-friendly
bring-your-own LLM endpoint
/ 04 · everything in the box

No Frankenstack.
No half-built glue.

5 categories · 14 tools
Runtime
  • Bun
    Runtime + dev server
  • Vite
    Prod builds
  • TypeScript
    Strict mode
Backend
  • Convex
    Reactive DB + fns
  • Better Auth
    Auth, sessions
  • Zod
    Runtime types
Frontend
  • React 19
    Concurrent
  • React Router 7
    Data routing
  • shadcn/ui
    Components
  • Tailwind v4
    Zero-config
SaaS
  • Autumn
    Stripe, 10× faster
  • Resend
    Outbound email
  • Lingui
    i18n full-stack
Quality
  • Biome
    Lint + format
  • Tests
    Preconfigured
  • CI
    GitHub Actions
/ 05 · pricing

Pay once.
Ship forever.

One-time license per use case. No seats. No usage tiers. Source included.

Tier 01 · Founder
$399 one-time

For your own products.

  • Full source code, forever
  • Complete AI Module (streaming, queue, boundaries, agents, tasks)
  • Unlimited projects you own
  • Lifetime updates
  • Circle community access
Use for: your SaaS, your side projects, your startup.
Not for: client projects you bill for.
Buy Founder — $399
/ 06 · questions

Skeptical? Good.
Answers below.

Why not just vibecode it myself?

100M+ tokens burned on GPT-5.4 XHIGH and Opus 4.6/4.7 MAX. Months of carefully designed modules and security hardening, backed by 20 years of coding.

You will not one-shot this quality. You'll spend weeks on an inferior result — which is exactly why I built RED. The baseline here is first-grade SaaS craft: the kind you see in Notion, Airtable, Linear.

Skip the burden. Push the RED button.

I'm not sure my product is agentic. Should I still buy RED?

If your product runs on an LLM today — or might in the next 18 months — yes. Every SaaS is becoming agentic, it's a matter of when, not if. RED isn't a SaaS kit with AI bolted on. It's a kit where the AI Module is the spine: auth, billing, i18n, routing all wrap around it. You get the primitives (streaming, queue, boundaries, async tasks) exactly when you need them.

Why is the AI Module built on Convex specifically?
Convex's reactive queries are a free realtime layer. A message lands in the DB, every subscribed client updates. No socket servers, no pub/sub infra — multi-party streaming for free.
What's actually different about external (async) Tasks?
Sync tasks run inside the chat turn — quick tool calls, under 5 minutes. External tasks run off-thread with unlimited runtime and push progress + completion back to the originating thread via websocket. The user can close the tab and get a notification when it's done.
How do Boundaries work in practice?
A Boundary is a project container with members, permissions, a scoped knowledge base, and its own memory. Threads inside share everything; threads in a different Boundary can't see in.
Is this just a Next.js template with a wrapper around OpenAI?
No — and it's not Next.js. RED is React 19 + React Router 7 + Convex. The AI Module is original engineering on top of Convex: agent abstractions, queue primitives, boundary model, task runtime. Works with any model provider (OpenAI, Anthropic, Groq, local).
Can I try it before I buy?
Yes. The live demo signs you in as an administrator with access to everything. The database wipes every 4 hours so you can break anything you want.
What's the license?
One-time purchase per use case. Ship unlimited SaaS products on top. You can't resell the kit itself. Lifetime updates included.
Refund policy?
14 days, no questions. If RED doesn't save you a month of work, email us and we'll refund.
/ 07 · last thing

The boring part
is already done.
Go build the product.

Every hour wiring websockets is an hour your competitor spent talking to customers.