# Emomo

> An Agent-first meme search CLI and skill. Agents reason about intent and inspect images; Emomo retrieves complete existing images.

- Website: https://timmyagentic.si/
- Docs: https://timmyagentic.si/docs
- Source: https://github.com/timmyagentic/emomo
- Release: https://github.com/timmyagentic/emomo/releases/tag/v1.0.0-beta1
- API: https://api.timmyagentic.si/agent/v1
- Skill: https://timmyagentic.si/SKILL.md

One-prompt installation: read https://timmyagentic.si/install.md and perform its installation and verification steps.

Manual installation with Node.js >=22.13:
npm install -g --ignore-scripts https://github.com/timmyagentic/emomo/releases/download/v1.0.0-beta1/timmyagentic-emomo-cli-1.0.0-beta1.tgz
emomo skill install --agent codex

Use emomo search "开心" --limit 5, inspect actual candidates, then emomo download <returned-id> --dir <new-output-directory>.

The shared keyword service has 7,302 static primary images. No model inference, embedding, automatic OCR, query rewriting or legacy fallback. No model credentials required. Approximate rate limit: 30 requests/minute/IP per Cloudflare location, eventually consistent; not a global hard cap. Follow retryAfterSeconds. HTTP search uses canonical protobuf REST JSON. Browser Origin requests are rejected. Local catalogs additionally support animations and subject/intent filters; never assume a saved local catalog is the public gallery. A missing local disk does not cause remote fallback.

Code is MIT. Image rights remain with their owners; code licensing does not license the images for commercial use.