Run visual Techniques from a conversation

FLORA's MCP server lets an assistant such as Claude or Codex discover and run the visual Techniques a team has built in FLORA. From the same conversation, the assistant can pass a brief, use reference assets, ask for variations and branch from selected results. Generated work remains available on the FLORA canvas for visual review and further editing. FLORA announced MCP in May 2026, so this is an established workflow capability rather than a fresh launch.

Setup uses the FLORA account through an OAuth sign-in on the first tool call. MCP and the API use the same Techniques and workspace billing; moving the prompt into an assistant does not make generations free. The documentation shows conversational requests such as running a Technique for a grid of options, then taking two favorites into another variation pass.

Example: explore an independent cinema campaign

Consider a small cinema promoting a season of restored films. Its designer could define a Technique for moody poster imagery, give the assistant an approved visual brief and references, and request a few distinct directions. After reviewing the results on the canvas, the designer could ask for variations of the strongest two. FLORA's canvas supports connected image workflows, so those alternatives can remain visible as part of one exploration rather than disappearing into separate chats.

That is a proposed use case, not a reported FLORA or XANTRANS campaign. A designer would still check whether an image suits the film and audience, select the final direction, and place it in an editable poster layout with approved title, credits and dates. The destination's dimensions and production requirements should guide that layout. MCP supplies candidates and iteration; the campaign's final composition still requires design judgment.

Sources & verification

Prepared with AI assistance from FLORA's official documentation. The cinema campaign is an editorial example, not a measured result or XANTRANS test.