About
/dev/fast builds an AI-native code forge for developers and engineering teams, including AI-assisted code review, agent trace storage, and token-cost optimization. Its differentiation is a focus on reducing review time and token spend while keeping engineering context tied to version control and enabling secure deployment in customers’ own VPCs.
Market
/dev/fast competes in AI-native software development infrastructure, spanning automated code review, AI-agent observability and cost control, and next-generation Git hosting. Its positioning is broader and more infrastructure-oriented than add-on reviewers such as CodeRabbit and Qodo: it combines architecture-focused review with version-control-linked agent traces, token optimization, VPC deployment, and an object-storage-based Git server. Compared with GitHub and GitLab, its differentiation is a purpose-built AI-native forge aimed at improving review throughput and reducing the operating cost of AI-generated code.
Software developers and engineering teams building with AI coding agents, particularly security-conscious organizations that want code, agent traces, and related infrastructure hosted in their own VPC. Likely buyers and users include developers, technical leads, and engineering leaders responsible for code quality, review velocity, and infrastructure costs.
At a Glance
Problem
/dev/fast addresses the bottleneck created by AI-assisted software development: generating code is becoming cheap, but reviewing it and exercising engineering judgment remain expensive. As AI increases code volume, review workloads can grow faster than teams can absorb them, while developers also face rising model-token costs. The company’s stated economic value is a 2x reduction in mean time spent in review alongside token-cost savings.
The principal use case is large pull-request review. /dev/fast claims it can let engineers review PRs ten times the size in half the time, shifting attention away from implementation details and toward architecture, correctness, and higher-value judgment.
Product / Service
/dev/fast positions itself as an AI-native code forge rather than a standalone coding assistant. Its initial review product, listed as being in alpha, is designed to analyze substantially larger PRs and help developers focus on architectural concerns. Supporting infrastructure includes queryable records of agent runs tied to version control, token-spend analysis and optimization, and a Git-compatible code server rebuilt on object storage.
The delivery model is aimed at engineering teams that need security and operational control: the company says its trace-storage infrastructure can run in a customer’s own VPC, while its code-storage and token products are marked private. The intended benefit is an integrated development environment for AI-generated code that combines faster review, lower model spend, durable engineering context, and more reliable source-code storage.
Market
/dev/fast competes in developer tools at the intersection of version control, AI-native software development, and automated code review. Its conceptual incumbent is GitHub—the company describes itself as building an AI-native version of the GitHub model—while adjacent AI-review competitors include CodeRabbit, GitHub Copilot Code Review, Sourcery, and Ellipsis. Broader AI coding platforms such as Cursor, Claude Code, and Amazon Q Developer also compete for developers’ AI-development workflows, although they are not identical products.
The company appears to be at an early commercial stage rather than showing established public traction. Y Combinator lists it as founded in 2025, part of Winter 2026, active, and based in San Francisco with a four-person team; the company site labels review as alpha and several infrastructure products as private. No public customer count, revenue, or usage figures were found, so it is best characterized as an early-stage, likely pre-revenue or limited-release startup rather than a company with demonstrated commercial scale.
Founders & Leadership
Funding History
Y Combinator
Recent News
/dev/fast announced Carson, an AI product that generates dynamic interfaces such as research dashboards and slide editors tailored to the user’s task. The announcement highlights enterprise-grade security and natural-language workflows for sales, marketing, and operations.
Y Combinator profiled /dev/fast as an AI-native code forge. The company says developers use it to reduce token costs and cut mean time in review by 2x; it is listed as an active Winter 2026 company.
Active Roles
0No active roles right now.
Get notified when they postBusiness Model
Its best-supported model is paid B2B software access for engineering teams, likely through subscriptions or enterprise licensing for its code-forge and secure infrastructure products. Exact pricing is not publicly disclosed; an associated Carson product offered a seven-day free trial, suggesting conversion from free access to paid usage.