Editor extensions · Augment
AI coding agent with deep codebase context.
Release history
We've got big news.Cosmos, Auggie CLI, the Code Context Engine, and related technology are joining Harness. We've agreed to sell these select assets because Harness's integrated platform delivers these capabilities to customers more effectively than building…
TL;DR: We built a software factory to automate mechanical work across the SDLC. Over eight months, size-adjusted output per developer grew 4.5× while merge times and revert rates fell. We added agents wherever work was piling up, rather than automating in…
Most agent platforms provide building blocks and leave the customer or a Forward Deployed Engineer to assemble them. Cosmos carries the implementation expertise inside the platform and can perform its own adoption work. Advisor turns years of agent engineering…
TL;DRAs our two-engineer Cosmos Advisor team shipped more automations to more customers, weekly product feedback grew rapidly - reaching 30+ feedback-threads per week. Investigating questions, reproducing bugs, finding owners, filing tickets, and putting up…
Every token you spend should buy as much finished work as possible is one of the core principles behind Cosmos and the Auggie CLI. It drives engineering decisions long before it shows up in anything customer-facing. There are two levers on value per token. One…
TL;DRIn an AI-native engineering organization, generating code is no longer the slowest part of shipping software. Review, verification, repair, and human decision-making become the bottlenecks, and ultimately define how much faster the organization can move.…
I lead the Solutions Architecture team at Augment Code. My team supports large engineering organizations evaluating Cosmos, our agentic orchestration platform, as the software factory their teams build in.A question I get often is: "Mayur, how healthy are our…
In April, Matt wrote that the era of single-model engineering is over. Three models cleared the bar at the time, and he predicted there would be more by the end of the year - both better and cheaper.It took just a few weeks for him to be proved right.What just…
“Loop engineering” is the practice of designing agentic loops: systems that take work from a trigger through execution and verification to a finished outcome, with agents doing the middle and humans at the checkpoints where judgment matters. Teams run them…
Internally, a significant amount of engineering time is consumed by manually verifying that the code an agent writes works E2E in dev. We sought to change that.We built an internal agent, the Verifier, to take over that step. On every run it deploys the change…
At Augment Code, like at many other companies, the human data team was the bridge to the data. Every question came through us. The trouble with being the bridge is that you're also the bottleneck, and ours backed up constantly. Context got lost jumping between…
TL;DR. Agents can now write code, review PRs, and investigate incidents. The next bottleneck is the large project itself: the design, decomposition, and coordination work that happens around the PRs. We built Project Builder, a Cosmos expert that turns a…
Last month I opened a roughly 1,000-line PR. The colleague I asked to review it looked at it and said, “Okay, give me 45 minutes. I’ll do it now.”A year ago, that same engineer would have said, “This is six or seven hours. I’ll get to it next Tuesday.”What…
Modern engineering teams want agents working everywhere across the software development lifecycle, not just in their IDEs. They want agents responding to issues, reviewing code, fixing builds, and triaging incidents. They are ready for a new kind of…
As coding and code review become increasingly automated in AI-native organizations, one of the next major bottlenecks is incident management. On-call engineers are often unable to contribute meaningfully to feature development during their rotation, and in…
TL;DR: We benchmarked Auggie vs Claude Code on Opus 4.7. Auggie takes a modest lead in quality (67.4% vs 66.3% pass rate) while costing ~33% less, thanks to sharper retrieval that results in token efficiency.Augment’s Context Engine was built to deliver…
TL;DR We surveyed 219 engineering leaders. 48% of their code is now AI-generated. 55% are worried about losing shared understanding of their own codebase. Asked for three words to describe how they feel, many put contradictory emotions in the same breath:…
TL;DR. When agents write 100% of your code, the bottleneck moves to review: at one point we had 1,400 open PRs and a 20-hour median time-to-first-comment. We rebuilt the review process on Cosmos as a team of agents that auto-approves low-risk changes, runs…
Today we're putting Augment Cosmos into public preview. Cosmos is the operating system for agentic software development: agents that run anywhere (your env, our cloud) and work across your SDLC, humans steering where judgment matters. And, Cosmos is built for…
TL;DR: Prism is a new option in the Augment model picker that routes each user turn to whichever underlying model best fits the work. For a team sending 10,000 user requests per month, Prism translates to roughly $20,000 in monthly savings, or up to 30% lower…
We added Karpathy-inspired coding rules to root AGENTS.md and ran three coding agents through 40 OpenClaw PRs. The judge scored code quality as basically unchanged, but the agents got to the same answers with less work: fewer tool calls and file reads, and…
Every engineering leader I talk to is trying to transform. They want their teams to be AI-native. They're rolling out tools, tracking adoption, watching dashboards. But the harder questions are the ones most leaders aren't asking out loud. What are we actually…
We pulled dozens of AGENTS.md files from across our monorepo and measured their effect on code generation. The best ones gave our coding agent a quality jump equivalent to upgrading from Haiku to Opus. The worst ones made the output worse than having no…
Thirteen months ago, I wrote a post arguing that model pickers were a design failure: the tool should pick the model. Developers should focus on shipping.This month, we added Gemini 3.1 Pro to Augment Code, our third model, alongside Claude and GPT-5.The…
What should my engineering org look like in 12 months? That's the question underneath every AI tool evaluation, every POC, every seat license debate. And it makes sense: the first step to provisioning a team is to understand what their day-to-day looks like…
For a decade, my professional identity was designer-who-can-code. I was in VS Code more often than Figma. My deliverables were often functional prototypes. “Here’s how it would feel to use it.”But now? Every designer can ship functional prototypes with their…
Starting today, Gemini 3.1 Pro is available as a model option in Augment.We ran it head-to-head against Opus 4.6 on the same task: plan out adding a new field end-to-end across the codebase. Gemini caught a subtlety Opus missed. That got our attention.Gemini…
Code review is broken. Not the tooling. Not the AI. The format.Every engineering leader we talk to says the same thing: their team ships fast, their AI catches real bugs, their static analysis is dialed in. And yet PRs sit in inboxes for days. The bottleneck…
Everyone in our industry is talking about AI and coding. How many lines can agents write? How many PRs can they generate? How fast can they ship? But if you pop out one level, the question that matters is different: what does coding quickly actually give…
Every developer has a workflow they've spent years refining.Mine lives almost entirely in the terminal. I'm SSH'd into a development machine. I organize my sessions with tmux, edit code in Vim, and use git worktrees to juggle multiple tasks in parallel.When…
Your team merged 30 PRs yesterday. How many did you actually read?Not skim. Read. Understood the tradeoffs, checked the edge cases, thought about how the change interacts with code three directories away.If you're honest, the number is small. Maybe two or…
In traditional development, specs were mostly planning documents. Once implementation started, they quickly became stale as the reality of the system emerged during coding.In AI-native workflows, the relationship flips. Specs become living contracts that guide…
How do you hire engineers when agents write 99% of the code?We found ourselves asking that question after realizing our hiring process was built for a different world. Software engineers have always done more than just write code, but the ability to write code…
The only durable moat in AI code review is review quality.Features, UX, and pricing matter — but none of them matter if developers don’t trust the feedback. If a tool produces noisy comments or misses real bugs, engineers quickly learn to ignore it.We believe…
Watch a painter work and you'll notice something obvious. They don't just paint. They step in, make a stroke, step back. Look at the whole canvas. Think. Step in again.The doing and the thinking are both part of the work. Neither one is the interruption.A…
I've spent my career in analytics engineering, and the most important aspect of the work I do isn't what most people think. It's not the dbt models, dashboards, or even the metrics themselves. The most important thing I build is business context around the…
You’ve been at it since 8. You chugged your morning coffee, you set up the perfect eval harness, and you wrote the most beautiful system prompt. You tested and refined almost every word. It’s free of contradictions. It uses words like essential. Your eval…
GPT-5.4 is OpenAI’s newest model, and it’s the first one we’ve used that feels built for agent workflows: planning cleanly, delegating well, and consistently following through without getting lost halfway. That matters for how we build and ship in Augment, so…
The question is no longer whether AI works. It’s whether your team can transform in time to use it.Something shifted in November. I've spent the months since in back-to-back conversations with customers, partners, and engineering leaders, and the same thing…
There's a GitHub repo for sharing AI coding agent rules. It has 37,800 stars. It has 68 contributors.That's a 556-to-1 ratio. For every person who contributed a rule, 556 people copied one without reading it.This is the state of AI agent configuration in 2026.…