Derneliche

[stage @stagedhappen]

DotCode has processed 46.6M tokens since launch.

We've also reduced DotCode's avg error rate to just 2%, following a series of runtime and reliability improvements that have now been deployed, with more upgrades coming soon.

We will be integrating a public facing analytics page tomorrow, so users can track all of this information in real time.

[Dot @usedotai]

DotCode is now officially live and free for the next 48 hours.

Generate applications, websites, dashboards, internal tools and full-stack software inside a private, isolated runtime built for AI-native software engineering.

Build anything, leave zero trace: https://t.co/eFFR0f09oV

[stage @stagedhappen]

DotCode is now officially live and free for the next 48 hours.

Generate applications, websites, dashboards, internal tools and full-stack software inside a private, isolated runtime built for AI-native software engineering.

Build anything, leave zero trace: https://t.co/QIRgoLbJdy

[Kazi @TheCryptoKazi]

@stagedhappen @usedotai Bro do you ever sleep

[stage @stagedhappen]

DotCode is now less than 24 hours away from launch, and I wanted to share some reflections on Dot, and the industry:

The Dot ecosystem has been publicly monetized for just over two weeks. During that time we've processed 44M+ tokens (110M+ since inception), generated $5.8k+ in inference credits, and mechanically settled and burned $1.75k+ in native $DOT.

We secured our first public enterprise API customer, completed another private enterprise onboarding, and will be announcing a significantly larger enterprise deployment on Tuesday.

These numbers showcase true appetite for what we are building, and the value users and enterprises are able to extract from our ecosystem.

To some, our usage may look negligible, irrelevant and insignificant. And when comparing Dot to the biggest AI providers in the industry, this comparison absolutely makes sense.

But for a platform that launched in monetized access 2 weeks ago, started at zero, operates in an emerging market, and is yet to drop it’s biggest product, this is a monumental feat.

Now, I’m sure, there are some who may still view our data as small, regardless of competitor analysis. And this is, unfortunately, a consequence of an industry standard that has made data inflation a common practice.

So much so, that it’s fostering an environment where newer protocols feel pressured to overstate their adoption simply to look competitive, continuing this damaging cycle.

At Dot, we will never conform to this flawed ideology, as it is simply not the route to building a long lasting, sustainable platform. Our data is shared for one singular reason; to openly and honestly measure growth with our community, as we continue to capture market share and grow platform revenue.

If you are, as a founder, overly concerned with aligning your business around the expectations of others, then you will naturally strip yourself of any defining or unique characteristics, real progress or genuine innovation.

Preoccupation with appearance over substance is a failure of mindset, execution, and foresight.

At Dot, we are acutely aware of this fact, and actively seek, in everything we do, to define this industry, not simply align with it.

All we have, and all we will ever have, is an uncompromising commitment to building great products, earning trust through transparency, and executing relentlessly on our vision.

Dot.

[stage @stagedhappen]

@usedotai Frontier grade agentic coding. Fully private. Objectively better AI.

24hours.

[Dot @usedotai]

One prompt - $0.50 - Zero trace.

A private, browser-native coding sandbox designed for frontier-grade agentic software engineering.

See the world differently.

DotCode. 24 hours.

[stage @stagedhappen]

Fable-5 no longer able to code? Not a problem with Dot Supercharged, our fable-5 level fusion model.

Check out the benchmark on this exciting model below ↓

[Dot @usedotai]

Rumors are circulating that Fable 5's coding capabilities have regressed.

Dot users do not need to worry with Dot Supercharged, our Fable-5 level fusion model.

Available at: https://t.co/eFFR0eZBzn

Dot Supercharged is a fusion architecture designed to deliver Fable-5 level reasoning without depending on any single model.

Check out our demo, where we challenged Dot Supercharged with one of the most demanding software engineering prompts we could construct. It produced a complete system architecture, implementation roadmap, and technical rationale with ease.

While the industry debates the capabilities of individual frontier models, Dot is building the private orchestration layer above them; combining routing, drafting, verification, and privacy into a single inference pipeline.

[stage @stagedhappen]

We have spent $20,000 on dedicated 8x NVIDIA H100 compute for DotCode.

We've signed a dedicated GPU capacity agreement with our provider, including a DPA/no-training commitment, for an 8x H100 DotCode runtime at $20k/month.

DotCode will not be an MVP, but rather, a fully fledged, agentic coding environment. A consumer application built from the ground up around private AI infrastructure.

This requires serious compute, and a strict, privacy-first architecture. We do not want coding agents and user sandboxes depending on random third-party model APIs.

Additionally, we are bringing a dedicated coding model to this infrastructure, giving users a stronger DotCode experience at a more competitive price.

8x H100 80GB HBM3
640GB GPU VRAM
NVLink topology
~1.5TiB system RAM

This lets us run Dot-managed coding infrastructure for our users: private sandboxes, strict log minimization, no training on user code, and a clear infrastructure agreement with our GPU capacity partner.

[stage @stagedhappen]

The new Dot website is now live.

We have shifted towards a more product-focused experience, making it easier for new users to understand exactly what Dot offers and how to get started with our growing suite of tools.

Try it now: https://t.co/3xsDk2jW3V

A major focus moving forward is improving the overall user experience, building a more polished, intuitive, and user-friendly platform designed to make adoption easier, ensuring frictionless entry into our ecosystem.

Over the coming week, we will continue refining the website, alongside a full documentation refresh to better showcase our expanding platform, products, and capabilities.

Forward.

[stage @stagedhappen]

We seek to change this industry, indefinitely.

7.

[stage @stagedhappen]

DotCode, our biggest ambition to date, is live in 7 days.

A fully private, sandboxed coding agent that gives you a safe environment for you and your projects.

No npm install. No exposing your workspace. No sacrificing privacy.

Build, iterate, and create inside a controlled environment designed for agentic software development.

This is the moment we've been building towards.
This is the moment that changes everything.

[Derneliche @Derneliche]

@usedotai Yes, and yes and yes, DOT going for all of it

[Dot @usedotai]

DotCode.

An industry first, fully private, agentic coding consumer application.

Turn intent into software, with autonomous code generation, tool execution, asset creation, and verification inside a secure sandbox.

Build. Iterate. Create.

7 days.

[Alaa @AlaaDelRey]

GitHub MCP is powerful, but raw tools aren’t the full product, so we added what's missing.

In Dot, we wrapped it with private OAuth, read-only defaults, account-aware repo context, model routing, and permission controls so users can ask “what’s on my GitHub?” naturally.

Frontier-style connectors, but private.

[stage @stagedhappen]

We have secured our first enterprise contract through @bloom_terminal, with more being finalized as we type.

Bloom has onboarded with a rolling $300 credit commitment and is now utilizing Dot API through our enterprise package, receiving a 30% discount on inference.

This marks the beginning of enterprise adoption for Dot, bringing private AI inference into real-world workflows at highly competitive rates.

Bloom now has access to powerful AI infrastructure through Dot’s privacy boundary, with enterprise-grade access at one of the most cost-efficient rates in the market.

[Derneliche @Derneliche]

@usedotai @usebloom_base

[Dot @usedotai]

We have secured our first enterprise API customer, @usebloom_base, with additional enterprise customers currently being finalized.

Bloom is now using our API to power private AI inference inside its workflows, accessing Dot’s privacy-first inference infrastructure through our enterprise package.

This marks an important milestone for Dot, moving from consumer usage into real enterprise deployments.

As AI becomes embedded into increasingly sensitive workflows, enterprises need infrastructure that provides both capability and privacy. Dot is providing that infrastructure.

[stage @stagedhappen]

Some thoughts I’ve had tonight:

We launched Dot 1 month ago. The Dot platform went live in public, monetized access 7 days ago.

Since then we have already seen 23 million tokens spent and $1.6k in inference credited.

Since enabling $DOT inference spend 3 hours ago, we’ve seen $170 in credits settled in our native token.

In comparison to mainstream providers, and our biggest competitors, this is tiny, marginal, insignificant.

But for a platform that started at 0, has been live for 7 days, and is being developed by a team of 3, this is monumental.

It is a real testament to the value users are extracting from our platform, and a reinforcement of our vision.

Decentralized AI inference is entering the mainstream whilst regulation and US enforcement is quietly increasing. These signs point to Dot quickly becoming a necessity for our users.

And the best part? DotCode is live very, very soon. A fully agentic platform that will revolutionize Dot as a service provider.

When all is said and done, we look to position ourselves as number 1.

Forward.