Locofy
Last reviewed
Jun 4, 2026
Sources
20 citations
Review status
Source-backed
Revision
v1 ยท 2,166 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 4, 2026
Sources
20 citations
Review status
Source-backed
Revision
v1 ยท 2,166 words
Add missing citations, update stale details, or suggest a clearer explanation.
Locofy (stylized Locofy.ai) is a Singapore-based design-to-code platform that converts user-interface designs from Figma and Adobe XD into production-ready frontend code. The company was founded in 2021 by Honey Mittal (CEO) and Sohaib Muhammad (CTO), and it ships a plugin that turns design files into framework code for React, React Native, HTML/CSS, Next.js, Gatsby, and Vue. In January 2024 Locofy launched Locofy Lightning, a one-click conversion tool built on what the company calls Large Design Models (LDMs), a class of AI models trained on visual design data rather than primarily on text. Locofy positions itself in the design-to-code segment alongside competitors such as Anima, Builder.io's Visual Copilot, and TeleportHQ. By mid-2023 the company had raised roughly $7.3 million across two early funding rounds backed by investors including Accel, January Capital, Golden Gate Ventures, and Northstar Ventures.
Locofy was started in 2021 by Honey Mittal and Sohaib Muhammad, two technologists who had worked together for more than a decade at regional Asian technology companies. Mittal studied computer science at the National University of Singapore on a scholarship, interned at Microsoft, and went on to serve as a product leader (chief product officer) at the startups Homage, FinAccel (Kredivo), and Wego. Muhammad was an engineering lead at companies including Wego, Homage, and Gumi, and the two had collaborated across travel, healthcare, and developer-tooling problems for years before founding the company. Both founders are first-generation immigrants to Singapore, from India and Pakistan respectively.
By the founders' own account, the idea came when Muhammad grew frustrated at the prospect of hand-coding hundreds of screens for a product revamp and proposed automating the conversion of designs into UI code. The company was registered, validated with founder friends, and raised an initial roughly $1 million over a single weekend, according to interviews Mittal gave later. Locofy launched a closed beta on 25 December 2021 ("Christmas night"), initially supporting Figma-to-React conversion, with early access for additional frameworks and tools to follow.
Locofy made its public debut on Product Hunt on 18 July 2022 and won that platform's Product of the Day, Product of the Week, and Product of the Month recognitions, according to the company. It later won Product of the Day again on the 2024 launch of Locofy Lightning.
Locofy raised two disclosed early-stage rounds. The structure and amounts below are cross-referenced across the lead investors' own announcements and multiple regional technology outlets. Note that some coverage rounds the cumulative total to about $7.5 million, while investor and database sources put it at roughly $7.3 million.
| Round | Date | Amount | Lead / key investors |
|---|---|---|---|
| Pre-Seed | March 2022 (closed 22 March 2022) | US$3.0 million | Accel and January Capital, with Golden Gate Ventures and Bold Capital, plus angels and operators |
| Seed (reported as "Seed II") | May 2023 (closed 16-17 May 2023) | US$4.25 million | Northstar Ventures (lead), with Golden Gate Ventures, Accel India, and Aviondor Group, plus founder-CTOs of Dropbox, LottieFiles, and Zopim |
The pre-seed round of US$3.0 million was announced in late March 2022. January Capital, one of the lead backers, described the financing as a pre-seed round and noted participation from Accel, Golden Gate Ventures, and Bold Capital, along with strategic angels who had founded or led companies such as Ola, Hasura, Holistics, Wego, and GajiGesa. The company said the capital would be used to expand its engineering and data-science teams.
The second round, US$4.25 million, was reported by the company and regional press as a "Seed II" round and closed in mid-May 2023. It was led by Northstar Ventures (a fund backed by TPG), with existing investors Golden Gate Ventures and Accel India and new investor Aviondor Group, plus a group of high-profile operator angels that press coverage identified as the co-founders and CTOs of Dropbox (Arash Ferdowsi), LottieFiles, and Zopim. The round brought Locofy's cumulative funding to roughly US$7.3 million. Around the same period the company added senior engineering and AI hires, including a former Wego CTO as VP of engineering.
Locofy converts a design file into frontend code through a plugin that runs inside the designer's tool (Figma or Adobe XD) plus a web-based environment the company calls Locofy Builder. The classic (non-Lightning) workflow has several stages:
Locofy Lightning, introduced in 2024, collapses much of the manual preparation into a single click. Instead of requiring the designer to manually optimize and tag the file, Lightning automatically performs design optimization, tagging, layer grouping, component detection, responsiveness, and class naming, and then produces frontend code. The company says Lightning automates close to 80% of the time-consuming parts of frontend development. Users can still review, preview, and fine-tune the generated code before exporting.
| Category | Supported |
|---|---|
| Design tools | Figma, Adobe XD |
| Output frameworks / languages | React, React Native, HTML/CSS, Next.js, Gatsby, Vue, Angular |
| Styling / extras | CSS, CSS Modules, TypeScript; Storybook export; GitHub sync |
Locofy's central technical claim is that design-to-code is better solved by models trained on visual design data than by general-purpose large language models. The company calls its approach Large Design Models (LDMs) and frames them as a design-domain analogue to LLMs: where an LLM is trained predominantly on text, an LDM is trained on design and layout data to reason about visual structure. CEO Honey Mittal drew the comparison directly to the rise of generative AI chatbots, saying that just as many people first grasped the potential of LLMs through tools like ChatGPT, Locofy aims to provide both the foundational design models and a product (Locofy Lightning) that makes them usable.
According to the company and contemporaneous technology press, the LDMs were trained on millions of designs and websites, and Locofy's stack combines multiple specialized models rather than a single network. Coverage of the launch described component sub-tasks (tagging, layer grouping, responsiveness, component detection, class naming) being handled by different techniques, including image-based neural networks, multimodal transformer models, graph-based neural networks, sequence-to-sequence models, stack-pointer networks, and heuristics, with conventional LLMs used for a minority of the pipeline. Technology outlets reporting on the launch cited the company's description of a unified model with close to half a billion parameters, and quoted Locofy's claim that its reliance on traditional LLMs was under 5% because the work is primarily visual rather than textual. The company has said it invested more than US$1 million in developing Locofy Lightning.
In July 2025 Locofy researchers published a technical paper, "LOCOFY Large Design Models: Design to code conversion solution," on arXiv (identifier 2507.16208, dated 22 July 2025), authored by Sohaib Muhammad, Ashwati Vipin, Karan Shetti, and Honey Mittal. The paper describes the LDM system in terms of a design optimizer (which uses proprietary ground-truth data to correct suboptimal designs), tagging and feature detection using pre-trained and fine-tuned models, and an "auto components" stage that extracts repeated UI structures into reusable modular code. The authors report that, in their experiments, LDMs outperformed LLMs on the accuracy of node positioning, responsiveness, and reproducibility, and they introduce a "preview match score" metric for evaluating design-to-code conversion accuracy. As with all single-company benchmarks, these comparative results are the authors' own and have not been independently reproduced.
Locofy offers a free tier and paid plans. The company has used a token-based model in which generation consumes "LDM tokens." Published pricing has included a free plan for basic design-to-code generation, live previews, and React/CSS exports; a pay-as-you-go option priced per LDM token; and Starter and Pro subscription tiers (offered monthly or annually) with progressively lower per-token rates and additional features such as API access, GitHub integration, team collaboration, and deployment webhooks. An enterprise plan with custom pricing and bulk token discounts is also offered. Exact prices have changed over time, so current figures should be checked on the company's pricing page.
Locofy has been covered by technology and startup press in both Asia and the West, including TechCrunch, which reported on the Lightning launch in January 2024, and outlets such as The New Stack and e27. The company states that builders have exported large volumes of code through the platform; in early 2022, investor January Capital noted that beta users had generated "millions of lines of code across hundreds of projects" with no marketing spend, and the company has more recently cited hundreds of millions of lines of code exported in aggregate (on the order of 366 million). These usage figures are self-reported and are not independently audited.
Independent reviews of the output have been mixed-to-positive. Reviewers generally praise Locofy for producing relatively clean, component-structured code with semantic HTML and consistent flexbox layouts, and credit it with meaningfully cutting initial layout work. Commonly noted limitations include occasional divergence from the original design (missing custom fonts or media), a need for manual refactoring and component deduplication, sometimes bulky generated CSS, and the fact that application logic (button behaviors, dynamic data, database integration) still has to be implemented by developers. Generating fully responsive code for desktop and mobile simultaneously has also been described as a challenge.
Locofy operates in the design-to-code (sometimes "Figma-to-code") category, a subset of broader AI code generation tooling. Its most frequently cited competitors are:
| Competitor | Positioning |
|---|---|
| Anima | A design-to-code tool aimed at designers who want code without learning a complex tool; reviewers describe its output as a literal, often heavily positioned translation of the design. |
| Builder.io (Visual Copilot) | A design-to-code plus visual CMS platform that maps generated code to existing design tokens and components; often positioned as the higher-end, enterprise-oriented option. |
| TeleportHQ | A visual builder that generates dependency-free code for React, Vue, Next.js, Angular, and HTML, but whose Figma workflow typically requires importing designs into its own editor first. |
Relative to these, Locofy emphasizes one-click conversion directly from the design tool, developer-friendly and editable component output, and its in-house LDM technology as a differentiator from approaches that lean primarily on general LLMs. The broader space also overlaps with prompt-to-UI and vibe coding tools, though those generate interfaces from natural-language prompts rather than from existing design files.