1
1 line
From one declared line,
Whether a person types it or an agent does, adding a field is one declaration — a name and a type. The database, API, screens and agent tools all come from that line, so it is the only line to review.
1name: field(String)
The TypeScript framework, agents included
Write one line. Deploy everywhere. Literally.
Web, iOS, Android, macOS, Windows and Linux. Your server and database. And the AI agents your users already talk to.
Screen → agent
The screen you build is the agent's interface.
Ask in plain words, and the agent works the same screen, in the user's own session.
Server → AI
Your server is already an MCP server.
Point any MCP client at your app, and it calls your endpoints through the same guards.
One code → every platform
Every platform, running at once.
The same app runs on iOS, Android, macOS and the web — save a file and every screen updates. Web pages render on the server, so SEO takes care of itself.
- Driven from Android; the rest follow
- Stack transition
- Bottom sheet
- Back, on every screen at once
- Fade transition
- Scale-out transition
- Stack transition
- A note, typed on Android…
- …saved to the database, live on every screen
- iOS
- Android
- Web (SSR)
- macOS
- Windows
- Linux
akan start · akan start-ios · akan start-android · akan start-desktop
All-in-one → scale-up
Server and database included, from prototype to IPO.
Most prototypes get rewritten once they have to scale. With Akan, the code you wrote on your laptop is the code serving at IPO — only the deployment grows.
Prototype
Running on day one
The server and the database come with akan start — nothing to install or wire.
Launch
One container to ship
The whole app, database included, deploys as a single container.
Growth
Add copies, not code
As traffic grows, more copies of the same app share it.
IPO
Many servers, same code
When one machine is not enough, it spreads across servers on Postgres and Redis.
name: field(String)One codebase, from the first commit to IPOPerformance
No rewrite for speed, either.
Database, SSR, auth and agents come built in, and it still keeps pace with the lightest Bun frameworks. It starts fast and idles light, so adding instances is quick and cheap.
123K
requests per second
102 ms
to start
57 MB
memory at rest
Requests per second
- raw Bun.serve138,072
- ElysiaJS137,780
- Hono129,926
- Akan.js123,319
- Fastify(Node)86,682
Apple M4 Pro MacBook Pro · production build · 50 concurrent usersFull benchmark
How it works
Single source of business
The screen an agent drives, the server any AI calls, the app on every platform — none of it is written twice. Everything is derived from one declaration like this, so nothing drifts out of step.
1name: field(String)
×8
8 layers
Implementation layers change with it
That one line reaches the schema, query, service, API, fetch, client type, state and UI prop — eight places that usually get edited one by one, now changing together. Nothing to chase, nothing to miss.
×6
6 platforms
Onto every platform
The same code ships as SEO-ready web, iOS and Android apps, and macOS, Windows and Linux desktop apps — with native-level screen transitions, not a wrapped website. One implementation to maintain, not six to keep in step.
×2
people & agents
For people and agents alike
Every guarded endpoint becomes an MCP tool and every control on screen an in-page agent tool, behind the same guards people pass. The screens and servers you saw above all start from that same line.
Agents build it, too
Maintainable, observable, and fast
With as little human in the loop as possible.
AI already writes good code. What slows it down is the person in the loop — answering its questions, checking the structure, reading the logs, prompting again. Akan settles those answers in rules and tooling, so the agent checks its own work and you only sign off on the result.
01
Maintainable
Conventions are lint rules, not a style guide. Every file lands where the rules say, so the codebase reads like one author wrote it — and a wrong turn fails the build before it reaches your review.
02
Observable
Every log line carries its request's trace, and a dev session writes one plain log file. The agent reads it, follows the trace and finds the cause — no logs to copy and paste for it.
03
Fast
A new field is one declaration, not eight edits, and a workflow MCP lets the agent plan, apply and validate a change on its own. You review the business change, not the wiring.
This is what we mean by agentic full-stack.
Code is how people and machines talk. Better AI doesn't mean you stop reading it; it means code must get easier to read.Agents do the writing. Akan keeps it short and consistent enough to take in at a glance.
From build to a live URL
Akan Cloud is the deploy platform built for Akan apps. Sign in from the CLI, share a preview, build, and ship it live.
Open Akan Cloud- 01
$ akan login
Sign in to Akan Cloud from your machine.
- 02
$ akan tunnel <app>
Share the app you are running on a public URL before you ship.
- 03
$ akan build <app>
Build the production artifact Akan Cloud runs.
The TypeScript framework, agents included
Read for humans.Write for agents.
Set up a new Akan.js workspace for me (https://akanjs.com). 1. Check that Bun 1.4 or newer is installed (`bun --version`). If it isn't, install or upgrade it as https://bun.sh describes. 2. Ask me for a workspace name and a first app name, short and lowercase, unless I wrote them below. 3. In this directory, run `bunx create-akan-workspace@latest <workspace> --app <app>`. It installs the akan CLI globally, creates ./<workspace> and installs its dependencies. 4. Inside ./<workspace>, start the dev server in the background with `akan start <app>` and check that the URL it prints (http://localhost:8282 by default) loads. 5. Tell me it's running, and that I should reopen you inside ./<workspace>: the project's Akan MCP server and AGENTS.md rules load from there.
