Read the screen
Find visible text, candidate surfaces, dividers, and color cues in the reference image.
Visual understanding for coding agents
Give your coding agent better eyes.
Kelra MCP turns interface screenshots into structured visual evidence—so an agent can build from what’s there, not what it guesses.
See how it worksEarly development · Starting with screenshot analysis
From pixels to a plan an agent can reason about.
Screenshot → structured evidenceA better starting point
See the interface.
Understand the evidence.
Kelra studies a reference screen and organizes what it can observe, what it can infer, and what remains uncertain.
Observe
Map how it fits together
Estimate geometry and relationships, while keeping confidence and uncertainty attached to the evidence.
Hand off a useful spec
Return a platform-neutral reconstruction spec that an external coding agent can use as its guide.
A tool in the workflow
Evidence in.
Implementation stays yours.
Kelra helps an agent understand a visual reference. The coding agent still owns the app code, rendering, corrections, and Git.
Evidence ledgerIllustrative
ObservedText and visual boundaries
InferredPossible spatial relationships
UncertainDetails the image cannot prove
Kelra MCP
Make the first pass
less of a guess.
A visual reasoning tool for the agents that build interfaces.