Framnova Guides · Product-team edited
From screenshotto useful context.
Start with the outcome: send a visual sequence to Web AI, attach implementation clues, or connect approved capture sets to a desktop coding client.
The guides are organized by trust boundary rather than feature count. Free Web AI delivery keeps every screenshot visible for review before you send it. Web Context adds only the selected element's bounded, sanitized implementation clues. Pro MCP either syncs one approved set to token-scoped hosted storage or exposes a local folder you chose. Each guide names the required permission, the data that moves, the storage boundary, the safest place to keep credentials, and the final verification step before you rely on the workflow.
Updated July 26, 2026Choose your workflow
05 PRACTICAL GUIDESConnect Framnova Direct MCP
Configure the endpoint and bearer token, sync one approved set, and verify the tools without committing a secret.
Capture selected webpage areas
Understand selected-region capture, ordered PNG originals, keyboard shortcuts, and the surfaces Chrome does not expose.
Send a screenshot sequence to ChatGPT
Arrange separate originals, prepare the attachments, and visibly verify every file before you press Send.
Add DOM and computed CSS context
Pair a screenshot with sanitized structure and relevant rendered style evidence—without claiming access to private source code.
Choose a screenshot MCP path
Compare a local read-only stdio workspace with authenticated Direct MCP over HTTPS, then inspect the shared tool surface.
Understand the evidence, not only the buttons
Read source-grounded articles about multi-image prompts, MCP trust boundaries, computed CSS, and reproducible UI debugging.