Alan AI

Coding agent writes code. Alan runs the lifecycle.

[Alan AI @tryAlanAI]

It's 5 AM on a Saturday and the team's still online because production decided it had other plans.

you don't build great companies on motivation. You build them with obsession.

If you're not obsessed with what you're building, you're probably not going to make it.

good morning @amannhq and @sidtalesara

[Shubh Srivastava @idleshubh]

@amannhq @steipete @tryAlanAI Are you still in office??

[Prasad Pilla @prasad_pilla]

@claudeai > fable too dangerous to release
> immediately releases new model better than it & nothing happens

[Alan AI @tryAlanAI]

@prasad_pilla @claudeai haha math is not mathingggg for real

[Alan AI @tryAlanAI]

@sidtalesara @amannhq @lowkeyarhann @prasad_pilla try it inside Alan https://t.co/83NKyufQVn

[Alan AI @tryAlanAI]

Opus 5 is live inside @tryAlanAI

Opus 5 approaches Fable level performance at half the cost.

Also it's the top performer in its family, up 22% over Claude Opus 4.7

team ships like crazy @sidtalesara @amannhq @lowkeyarhann @prasad_pilla

[Claude @claudeai]

Introducing Claude Opus 5.

It's a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price.

[Alan AI @tryAlanAI]

Opus 5 vs. Fable vs. Sol

[Claude @claudeai]

Introducing Claude Opus 5.

It's a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price.

[Alan AI @tryAlanAI]

@moonlitebyte for sure you can DM us

[Alan AI @tryAlanAI]

We built Whiteboards inside @tryAlanAI because we don’t want another disconnected canvas.

Yes, Miro exists. But Alan AI Whiteboards are built for the way our teams already work inside Alan: with live conversations, tasks, code, and decisions happening in one connected workflow.

For teams, that means less tool hopping, fewer handoff breakdowns, and less time rebuilding context from old threads. Product, design, and engineering can shape a plan in one space and move from thinking to doing without context drift.

When teams use the Alan AI Whiteboard, collaboration becomes more than visual, it becomes execution ready.

https://t.co/SL80w7U4gT today!

[Alan AI @tryAlanAI]

@gabriel1 yeah it did pretty good job in our landing page ngl

[Alan AI @tryAlanAI]

@aaliya_va exactly aaliya!

[Alan AI @tryAlanAI]

Single Model Loyalty Is the New Technical Debt.

Don’t be controlled by one model.

Orchestrate them all @tryAlanAI.

[Prasad Pilla @prasad_pilla]

I completely agree that the “one size fits all model” idea is already fading.

The real leverage is in orchestration: not replacing model choice with one champion model, but matching the right model to the right task, at the right time, with the right constraints (latency, cost, reliability, reasoning depth, style).

That’s exactly the direction we’re taking with @tryAlanAI building a model aware, task-aware stack where each layer can pick the best model for the specific job instead of forcing everything through a single endpoint.

It’s a more pragmatic and scalable way to build AI systems, and I’m excited to see this become the default mental model for product teams.

would absolutely love you to https://t.co/267w1GAgMy! DMs are open @hnshah

[Alan AI @tryAlanAI]

@Heyambarg @idleshubh @amannhq yesss sirrrr they don't let me sleep

[Alan AI @tryAlanAI]

@idleshubh @theo haha exactly https://t.co/83NKyufQVn to find what agentic coding is really about!

[Alan AI @tryAlanAI]

@theo tell me you haven't tried Alan without telling me you haven't tried Alan

[Prasad Pilla @prasad_pilla]

"I don't prompt Claude anymore. There are loops running, and it's the loops that prompt Claude and decide what to do. My job is to write the loops."

- @bcherny (creator of Claude Code)

That shift changes how you think about building with AI. A single loop can plan, execute, verify, and stop, but it quickly becomes a bottleneck.

The next step isn't simply adding more agents it's connecting specialized loops into a graph where work branches out, runs in parallel, verifies itself, and converges back into a finished outcome.

That's graph engineering. The teams pulling ahead aren't the ones with the most agents; they're the ones designing how those agents coordinate.

If you want a clearer picture of what graph engineering looks like in practice and why it's becoming the next evolution of agentic systems, read the article below.

[Alan AI @tryAlanAI]

That's exactly why we built Alan AI to orchestrate planning, implementation, reviews, and execution across engineering teams, so cheaper intelligence actually translates into faster shipping

https://t.co/83NKyufQVn today!

[Prasad Pilla @prasad_pilla]

Everyone is debating which model is smarter.

The bigger story is that intelligence is becoming absurdly cheap.

1M tokens went from $20 in 2022 to $0.07 in 2026.

2022: ██████████ $20.00
2023: █ $2.00
2024: ▍ $0.50
2025: ▎ $0.15
2026: ▏ $0.07

A 280× price collapse.

At some point, AI won't be constrained by inference cost. It'll be constrained by who knows how to use it.

[Alan AI @tryAlanAI]

@sidtalesara agentic board ftw!

[Alan AI @tryAlanAI]

@idleshubh @amannhq visualmaxxinggg lol

[Alan AI @tryAlanAI]

We've been betting on this for a while.

The endgame isn't another browser.

It's a superapp for knowledge work where browser, terminal, agents, memory, and models all live in the same context.

That's exactly the direction we're taking with Alan AI.

https://t.co/B4AsBHRV7b today!

[Amarnath @amrnth0]

Codex, Claude, Cursor are silently killing the browser giants like Chrome. The in-app browser on these apps makes it really convenient for knowledge workers who mostly live in their browsers. For people who are too afraid to migrate their logged in sessions to these in-app browsers, Chrome provides a native bridge through which Codex et. al. can interact with websites directly(using their extension).

Some founders are creating AI browsers, while the silent spark that burns down the forest could be Codex, Claude, Cursor which are approaching the problem from the other side - agent first, browser second.

All in all, the era of the superapp might be here.

[Alan AI @tryAlanAI]

@aashatwt INVEST EARLY IN AGENTIC ENGINEERING!!