We’ve officially launched Stela — the first public glimpse into the future we’re building. It’s just the beginning. Now, we’re on a mission to make it hellishly powerful. We’re actively pushing Stela toward its next stage of evolution — into something significantly more capable.

Our team is working in parallel on two major fronts — advancing Stela by unlocking the full potential of our in-house AI system and building a platform-agnostic AI coding assistant that works across diverse development environments beyond VS Code.

Together, these efforts form the backbone of what's next. Below, we share how we plan to advance Stela — from expanding its features to sharpening its intelligence — and how our in-house AI systems, Megaminds, will drive that transformation.

If this excites you, if you're passionate about creating products that redefine how software is built, and if you want to help shape that future with us — then you're more than welcome to join us on this remarkable journey.

A major upgrade coming to Stela is full codebase access — the ability to understand and operate over your entire repo, not just single files or visible tabs.

With full-repo awareness, it will learn your structure, dependencies, naming conventions, and file relationships — enabling intelligent, contextually accurate suggestions that fit your codebase like a native contributor.

This foundation of full-repo awareness will power Stela's ability to generate production-grade answers in chat — and will unlock the deeper intelligence running silently in the background, enabling smarter completions, structural understanding, and seamless refactors across your entire codebase.

One of the key upgrades we're building into Stela is a smarter, production-focused chat experience.

Today, most AI coding tools generate quick, shallow code that's full of issues — riddled with bugs, security flaws, and architectural mismatches. This low-quality code silently enters codebases and becomes a nightmare for testing, debugging, and QA.

We're fixing this at the foundation. Stela's chat is evolving into a place where high-quality, production-ready code is generated by default — aligned with your real architecture, conventions, and long-term goals. This leap will be made possible by Megamind 1, our in-house AI engine that will power the conversational experience inside Stela.

We're unlocking a whole new layer of intelligence in Stela — one that watches your codebase as you build.

In fast-moving teams, large volumes of code get added rapidly — and if that growth isn't grounded in the right structure, refactoring later becomes painful, costly, and often impossible to scale cleanly.

We're addressing this at its core. Stela will soon come with an embedded intelligence that continuously reviews your code as you write — quietly ensuring that every new line fits your architecture, flags vulnerabilities early, and keeps your codebase production-grade from the start. This capability will be made possible by Megamind 2, our in-house AI engine built to understand and evolve your codebase over time.

We're adding agentic capabilities to Stela — systems that go beyond just answering prompts. The next evolution of Stela will introduce autonomous agents that think in systems, not snippets. Agents that can reason through problems, plan multi-step changes, and take initiative to execute complex workflows. Soon, you'll ask Stela to refactor modules, write tests, or scaffold features — and watch it carry out those changes across the codebase with precision.

Today, most AI agents can't actually ship code. They produce fragile snippets riddled with bugs, integration gaps, and missing context. These agents may automate tasks, but they don't deliver real engineering value — and they often leave a mess behind.

We're taking a fundamentally different path. Our agents will be the most reliable autonomous systems for generating production-ready code — because we're already solving the hard part: powering and grounding every small code change with deep production intelligence. And these agents will be built on the same in-house AI systems — Megaminds. Megamind 1 writes with future demands in mind; Megamind 2 grounds every line in your actual codebase. Together, they power agents that will build production ready software — not just runnable demos.

One of the powerful capabilities we're adding to Stela is next edit prediction — the ability to anticipate what you're likely to change next, based on how your code and intent are evolving.

We're eliminating the repetitive, obvious edits that drain focus but add little value. By understanding your code's structure and your goals, Stela can proactively complete boilerplate, fix ripple effects, and adjust related functions or dependencies before you even ask — so you can stay in flow and focus on the high-impact logic that really matters.