AI Agent-Readable GIFs — Machine Layer Documentation

How AI agents read terminal GIFs through cast files, transcripts, .md endpoints, and structured APIs.

The Dual Layer

Every GIF on AgentGIF has two layers: a visual layer (GIF image) for humans and a machine layer (cast + transcript) for AI agents.

.md Endpoints

Append .md to any URL to get a text/markdown response optimized for LLMs:

curl https://agentgif.com/@user/demo/.md

Cast Data (asciinema v2)

GET /api/v1/gifs/<id>/cast/

Returns the raw asciinema v2 cast file — a JSON header + newline-delimited events with timestamps.

Transcript

GET /api/v1/gifs/<id>/transcript/

Returns a plain-text transcript of the terminal session, suitable for indexing and search.

Structured API

GET /api/v1/gifs/<id>/

Full JSON metadata including title, command, tags, author, embed codes, and download URLs.