/jump to repo

Repos Recent Bookmarks Watched Notes Tags Discover Compare Stats GitHub

Repository brief

abi/screenshot-to-code

Read the upstream summary on the left, browse the cached forks below it, and load each fork comparison into the right-hand panel.

Cached analysis

cached 2026-03-30T12:12:12.827Z

3mo ago

abi/screenshot-to-code

Open source AI tool for turning screenshots, mockups, and Figma designs into code. It is large and active, with 72,096 stars, 8,894 forks, and recent commits in March 2026. The repo includes a React/Vite frontend and FastAPI backend, so forks are likely most interesting for teams that want to extend the code generation pipeline, model support, or UI/workflow around screenshot-to-code.

Loading tags...

Stars72,096

Forks8,894

Default branchmain

Last pushed2026-03-26T13:01:38Z

Recommended shortcuts

Jump straight into Discofork's strongest cached fork picks, or open a compare view in one click.

Forks

Choose a fork to inspect

10 of 10 fork briefs

Maintenance:

Magnitude:

Sort:

Selected

Prefer this fork if you want a customized deployment-focused base and can tolerate lagging upstream. Prefer upstream if you want the newest model support, caching changes, and full evaluation/activity workflows.

Choose this fork only if its local prompt/eval and workflow changes match your needs and you are prepared to do ongoing upstream merge work. If you want the latest models and product capabilities, upstream is the safer choice.

Choose this fork only if you value its custom refactors more than upstream parity. For most adopters, upstream is the safer choice because this fork is stale and appears to have removed several newer workflows and UI areas.

Choose this fork only if you want a simplified, likely hosted screenshot-to-code experience and do not need upstream parity. If you want active maintenance, newer model support, or the full eval/agent workflow, upstream is the safer choice.

Prefer upstream unless you specifically want this fork's simpler, hosted-tool-oriented snapshot and URL screenshot workflow. The fork looks materially behind and appears to shed evaluation and broader product capabilities, so it is better as a lightweight starting point than as a current, full-featured base.

Choose this fork if you want a more experimental, developer-centric base for model and agent work. Avoid it if you want the newest upstream UI, model support, and polished workflows with minimal merge risk.

Choose this fork only if its eval/history/import workflow is the point of the project. For most adopters, upstream is the better default because this fork is stale, materially behind, and missing newer model and prototype capabilities.

Prefer this fork only if you specifically want its older eval/history workflow and are prepared to maintain a divergent codebase. If you want current model support, active upstream features, or an easier upgrade path, upstream is the better choice.

Choose this fork only if its evaluation- and agent-oriented changes are specifically what you need. If you want a current, broadly supported screenshot-to-code product, upstream is the safer default.

Choose this fork only if its custom model/default behavior, mock-LLM workflow, or simplified UI matches your needs and you are willing to absorb merge debt. If you want current upstream capabilities, active maintenance, and newer model/support work, upstream is the safer choice.

Fork comparison

liseami/screenshot-to-code

38/100

stale

significant_divergence

Prefer this fork if you want a customized deployment-focused base and can tolerate lagging upstream. Prefer upstream if you want the newest model support, caching changes, and full evaluation/activity workflows.

Likely purpose

Adapt the upstream screenshot-to-code app for custom deployment, local experimentation, and alternate model/tool workflows, likely with some UI simplification around evaluation and activity tracking.

Best for

Teams needing a customized self-hosted screenshot-to-code stack; Users who want to experiment with model plumbing or mock generation behavior; Adopters who value deployment-specific changes over staying current with upstream; Maintainers comfortable reconciling an older fork with a moving upstream

Additional features

Custom mock LLM and generation-path changes for offline or controlled testing
Updated backend provider/tooling work that appears to support custom model and request workflows
Deployment-oriented changes, including a commit focused on domestic deployment recommendations
Workflow and UI adaptations around code generation and evaluation rather than a plain upstream mirror
Adds plugin, tool, or MCP-style extensibility beyond the upstream baseline.
Introduces more operational workflow surface such as server-side handling, auth, session control, or admin tooling.

Missing features

It is 285 commits behind upstream, so it likely misses recent upstream model, caching, and request-settings improvements
Evaluation pages and related workflows appear to have been removed or heavily pared back (`BestOfNEvalsPage`, `RunEvalsPage`)
Agent activity visibility appears reduced because `AgentActivity` was deleted
Testing and build-related paths appear to have been trimmed or changed, which may make upstream-style validation harder to run
Some upstream behavior appears to be removed, disabled, or intentionally stripped down in this fork.
It trails upstream by 285 commits, so some recent upstream features and fixes are likely not present yet.

Strengths

More opinionated for customized deployment and experimentation
Potentially easier to adapt if you need to replace or fake model behavior
Large, mature upstream base with substantial existing code generation coverage

Risks

Stale relative to upstream's active March 2026 development
High likelihood of feature drift from upstream documentation and workflows
Removed evaluation/activity UI may make debugging and model comparison harder
The scale of deletions suggests the fork may be less suitable if you want full upstream parity

abi/screenshot-to-code · Discofork